As AI images flood the internet, can these tools help you tell what's real?
AI image detectors promise to help cut through the flood of online content to help determine what's real and what's not — but can you rely on them to get it right?
Many people now regularly turn to AI detectors to help determine if viral images are AI-generated. And as people increasingly become skeptical of even real images, those tools are in greater demand.
CBC News tested five commonly-recommended free AI image detectors to see if they were accurately able to estimate whether an image was real or AI-generated. These included: AI or Not, Illuminarty, Hive Moderation, Is it AI? and a tool hosted on the open source machine learning platform Hugging Face. While many of these AI detector services have a free tier, some have pricing plans that can cost hundreds of dollars each year.
The CBC's visual investigations unit ran three photos through each detector twice. The images tested include a real photo of a CBC hallway lined with lockers, a similar-looking AI-generated image, and a lower resolution, compressed version of the AI-generated image that was posted to the social media platform X.
Do AI image detectors work? We tested 5
19 hours ago
Duration 2:44
AI image detectors are growing in popularity as a way to determine whether an image or video shared online is real or not. CBC News' Visual Investigations team tested some of the most popular free tools online to see how effective they are — and whether you should rely on them.
In the first test, AI or Not and Hive accurately labelled all three images. Illuminarty and Is it AI? got all three wrong. The detector hosted on Hugging Face accurately labelled the AI-generated images, but thought the real image was fake.
In the second test, the results were the same, except this time Is it AI? labelled the real image correctly, but still got the AI-generated images wrong.
"It really does depend on which system you're looking at," said Ben Y. Zhao, a computer science professor at the University of Chicago.
Tests produced mixed results
CBC News reached out to all five AI detector companies for comment.
The CEOs of AI or Not and Hive both said their AI detectors are searching for patterns invisible to the naked eye.
"While some AI-generated images may be easier to identify by common errors — like extra fingers on someone's hand, unrealistic reflections and shadows, or the mismatched scale of objects — the sophistication of AI-generated imagery is moving beyond the threshold of human detection," said Hive CEO Kevin Guo in a statement.
AI or Not CEO Anatoly Kvitnitsky says their tool looks for "pixel-level patterns of content."
The creator of the AI detector on Hugging Face, developer Colby Brown, says AI detection is still worth pursuing, even if it gets some images wrong.
"User caution is needed," Brown said in a statement. "Individual images can fool such detectors even if they have reasonable accuracy on a larger sample (or feed) of images."
The team at Is It AI? said in a statement that the test CBC News performed highlights "the ongoing challenge that AI image detectors face" as the technology develops.
They also said that their tool "covers a wide range of domains and typically requires a larger and more diverse dataset to assess accuracy comprehensively."
As AI image generators are continuously improving, so are detectors. Is It AI? said the company will soon release a new detection tool with "substantial improvements" in accuracy. Brown also said that he may develop a new and more advanced tool.
Illuminarty didn't respond to CBC's requests for comment.
WATCH | Investigating the claim AI was used to boost crowd size at Carney event:
How we debunked this Carney AI claim
16 days ago
Duration 1:51
Zhao says some AI detectors are better than others.
"Some of them are trained on millions of images that allow them to do a better job with discerning the differences," he said.
He noted that bad actors can even use AI image detectors to iterate and fine-tune fake images that would then be labelled as real.
"I think the real danger is really to a lot of the folks who are not in a situation where they expect" to be targetted by AI-generated fakes, Zhao said, noting these are usually people who aren't as familiar with technology. "They're going to be easier targets for scammers and phishing scams and different kinds of things."
Zhao says that old tricks for detecting AI images are becoming less reliable. Famously, early iterations of AI image generators had trouble mimicking human hands, but he says that's not the case anymore.
Still, AI image generators don't get everything right, and a trained eye can often pick out details that clearly indicate AI was used.
The AI image CBC News used in the test can be identified as fake with the naked eye. The lockers in the hallway have locks that are warped and blurred, for example. The overhead lights have no fixtures, a panel on the ceiling has a line running through it, and there appear to be far too many lockers for the amount of space shown in the photo.
Zhao says that when people are trying to tell the difference between a real photo and an AI-generated image, thinking through the details is important.
"Does it make sense for the button to be placed in this way? Does it make sense for the hair to blend in with the turtle neck that way? Those kinds of smaller details are really still tricky for models to get right," Zhao said.
Methodology:
During testing, CBC News sought to mimic the experience of a member of the general public. We chose five free popular online AI image detectors. We chose five free popular online AI image detectors by tallying the number of recommendations from lists featured on the first five pages of Google search results and chose the top five of those services.
Three images were tested: a real photo taken by CBC reporters, an AI-generated image and a compressed version of the same AI image that was posted to X and downloaded again. The five detectors were then scored as correct or incorrect based on whether they accurately assessed whether the images were more likely created by a human or AI. The test was run twice.
The AI photo was generated from Google's Gemini AI with the prompt: "Create an image of a hallway with blue lockers filling half of the hallway on the left, grey checkered carpet and light orange wall on the right, and white hallway on the left back. A red wall is at the end of the hallway."
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Canada Standard
16 hours ago
- Canada Standard
Ottawa Battery Project Down But Not Out After Failed Committee Vote
A proposed 250-megawatt battery storage project in Ottawa's rural west is down but not out, after the city's Agriculture and Rural Affairs Committee (ARAC) voted unanimously last week to reject the plan. The Energy Mix has learned that supporters of the Marchurst Battery Energy Storage System (BESS) expect a closer margin when the full city council meets June 11-with the possibility that at least two ARAC members will change their votes. The 10-hour meeting last Thursday heard 68 public delegations before denying official backing for the project, CBC reports. Gatineau, Quebec-based Evolugen, a subsidiary of Brookfield Asset Management, already has a contract from Ontario's Independent Electricity System Operator to build the project on a 4.5-hectare site, about 30 kilometres west of downtown Ottawa. But it's conditional on a municipal support resolution (MSR) from the city. "The technology is a key part of the Ford government's plan to solve a looming energy supply crunch, as demand in the province is expected to increase by 75% by 2050," CBC writes. "But for many residents, the location of this particular battery trumped any broader provincial picture." Still, while the ARAC vote "was unanimous at face value," some committee members were "very conflicted", a committee observer told The Mix , with one councillor comparing the introduction of battery storage to the arrival of the motor car in a horse and buggy community. View our latest digests "I believe that BESS is a good technology, and this, on its merits, may actually be a good project," said Councillor Matthew Luloff (Orleans East-Cumberland), according to a segment of an unofficial meeting transcript viewed by The Energy Mix . While Luloff said his committee vote followed the wishes of councillor Clarke Kelly (West Carleton-March), who vehemently opposes the project in his ward, "I reserve the right to make my own decision at Council, where this matter will come forward again for further discussion." Luloff added: "I will be taking the lessons I learned here today at that vote. Let this also be a lesson on consultation, it is not a box to be checked but a fundamental part of decision-making. Community education is incredibly important, and we must continue to do better... in how we inform." Angela Keller-Herzog, executive director of Community Action for Environmental Sustainability (CAFES), said the community adjacent to the Marchurst site "needs to come together" in order to negotiate with Evolugen for local benefits from the project. But at the moment, "avowed opponents... distrust information offered up by the project proponent company," she wrote in an email. "The anti-BESS group is talking about 'raising pitchforks' and the local Facebook group ejects anyone who has positive things to say about the project." [Disclosure: CAFES is a partner of the Green Resilience Project, a joint community listening effort hosted by Energy Mix Productions and the Basic Income Canada Network.] The Energy Mix could not verify the content of the private West Carleton BESS Facebook group. Courtney Argue, a leading project opponent who lives about 400 metres from the site, said "it really depends" how visitors are treated when they join the group. "People get deleted if they are being disruptive or we have a suspicion that they have ties to Evolugen or their potential or confirmed partners," she told The Mix in an email. "You try your best. Sometimes you get it right and sometimes you get it wrong." As for the comment about pitchforks, "we hold our pitchforks in our hearts when it comes to protecting our land, wildlife, and way of life," Argue wrote. "Rural folk are not violent folk. We handle things together. We show up in the masses at town halls, we help neighbours. In this fight, we cannot leave any stone unturned." While the group is meeting neighbours who support sustainable energy solutions-Argue said she'd "love to add solar to our farm to help sustain our operations"-she maintained that most of the people she's heard from oppose the BESS project. Keller-Herzog said the local community newspaper, West Carleton Online, which has covered the issue extensively over the last two years, recently estimated that "the community was pretty evenly split but that the anti-BESS voices consistently tried to present a picture as everybody against." In spite of the furor and the "significant coverage" it has received, "there is probably a majority of busy working families that are not tuned in to the local news and have no engagement on the issue," she said. "The real problem? Our community lacks a calm and credible space to talk things through," Keller-Herzog wrote in a mid-May opinion piece for West Carleton Online. "Many thoughtful West Carleton residents, who might offer balance or ask good questions, have simply gone silent, understandably reluctant to wade into a conversation that has become conflictive and dominated by a few loud voices." Evolugen, meanwhile, "now appears hesitant to engage further, perhaps understandably, given how strident the misinformation has become," she said. Evolugen's Canadian head of development, Geoff Wright, had not responded to a request for comment as The Mix went to virtual press Monday evening. But "there are other BESS conversations that affirm of course residents and businesses in Ottawa want the lights to stay on, want the grid to be stable, want power to be affordable, want the jobs, investment, taxes paid and local community benefits from a safe, non-polluting energy infrastructure solution," Keller-Herzog added. Source: The Energy Mix
4 days ago
Faster isn't always better. Slow-charging EVs could have big benefits
When Julia McNally decided to buy an EV and started her research, she came across a lot of articles and ads pushing an apparent must-have accessory — a speedy home charger designed specifically for EVs. Everything was pointing me to Level 2, recalled McNally, director of climate action at Toronto Hydro. She knew that all EVs can do Level 1 slow charging, or trickle charging, from a regular 120-volt wall outlet, adding about six kilometres of range per hour (except in very cold winter temperatures, which can slow charging speeds). And she already had an outlet of those in her backyard, near the alleyway where she planned to park her new Mini EV. But more than four out of five U.S. EV owners used Level 2 for home-charging in 2023, according to market research firm J.D. Power (new window) . Using a higher 240 voltage, often needed for a stove or dryer, Level 2 chargers can add about 30 to 50 kilometres of range per hour and refill a typical EV's entire 400-kilometre range overnight. Meanwhile, Level 3, or DC fast chargers, often installed along major highways, can add 250 kilometres of range per hour (some are even faster (new window) ) and charge a battery to 80 per cent in 30 minutes. Get more with a free CBC account Comment on articles, stay in the know with our newsletters and stream more on CBC Gem. Sign In (new window) Create a free account (new window) Why faster may not be better Some experts, such as Daniel Breton, CEO of Electric Mobility Canada, have argued people "really need" Level 2 chargers at home (new window) , as it can take days to charge an empty battery to full at Level 1. But most people don't drive the hundreds of kilometres needed to empty their battery each day — and there's a downside to faster charging. You're adding cost, McNally said — potentially thousands of dollars. Installing a Level 2 charger requires a licensed electrician, she said. In Toronto, it means consulting with Toronto Hydro and the Electrical Safety Authority. And homeowners often will need to increase the size of their electrical panel, adding additional costs. But it's something more Canadians may be thinking about soon, amid Canada's zero-emission vehicle mandate (new window) , requiring that 20 per cent of cars, vans and light trucks sold in Canada be electric, hybrid or hydrogen-powered cars by next year. The goal is to reach 100 per cent zero-emission vehicle sales by 2035. Conservative Leader Pierre Poilievre had vowed to scrap the target if elected (new window) , but with a Liberal re-election (new window) , the target still stands. WATCH | Can northern power grids handle electric vehicles and heat?: Début du widget Widget. Passer le widget ? Fin du widget Widget. Retourner au début du widget ? Can northern power grids handle electric vehicles and heat? Electric vehicles and electric heating are expected to put a strain on northern power grids. A new study out of Yukon University offers utilities some ways to fix that. The CBC's Liny Lamberink has more. Level 2 charging isn't just more expensive and logistically difficult for individual EV owners. In some Canadian communities, the aging electrical grid may not be able to handle too much Level 2 charging at once. For example, a recent Yukon University study found that if more northerners install Level 2 chargers and electric heating, that could cause problems for transformers (new window) — a key piece of equipment in local electricity distribution networks. At the time of the study, published last December, there were only 88 EVs in all of Yukon, and half of them were plug-in hybrids. Blake Shaffer, a University of Calgary associate professor, studied the situation in his community with local utility Enmax. He previously told CBC News (new window) that electricity distribution networks would need significant upgrades in order for all EV drivers to be able to charge at Level 2. That's where the real challenge of EVs comes about, he said, noting high costs for both individuals and electric utilities. McNally says Toronto Hydro has adequate capacity for whatever EVs and heat pumps come at us. She acknowledged, however, that in cases where someone does ask Toronto Hydro for extra capacity you need to pay for the upgrades. Meanwhile, Level 1 takes advantage of wall outlets that people often already have, including residents of apartments or condo buildings. In colder parts of Canada, many parking spaces have a plug intended for block heaters. (Although tenants may have to negotiate with their landlord to use it for charging (new window) .) WATCH | This electric vehicle owner says tenants who pay hydro should be able to plug in: Début du widget Widget. Passer le widget ? Fin du widget Widget. Retourner au début du widget ? This electric vehicle owner says tenants who pay hydro should be able to plug in Renters might find themselves in uncharted legal territory if their landlord wants to make them pay for charging their electric vehicles — even if electricity is included in their lease. Many people don't need Level 2 at home Living in Toronto, McNally doesn't drive 400 kilometres a day; typically, she only covers 600 kilometres in an entire month. So she knew that Level 1 charging was probably good enough for her needs. That's not unusual — even outside Canada's largest city. Shaffer studied the driving and parking habits of 129 EV drivers in Calgary from December 2021 to December 2022. (While that was during the tail end of the pandemic, Statistics Canada reports very similar commute times in Calgary in 2022 and 2024 (new window) .) The study found 29 per cent of drivers only ever needed Level 1 charging (new window) because they drove very little relative to the time they were parked. Another 53 per cent could use Level 1 most of the time, but might need to visit a public Level 2 or fast charging station up to once per month to top off their battery. WATCH | Canada needs more charging stations to hit EV targets: Début du widget Widget. Passer le widget ? Fin du widget Widget. Retourner au début du widget ? Canada needs more charging stations to hit EV targets Experts say Canada needs hundreds of thousands more charging stations to support electric vehicle targets, but it's unclear who's in charge of building them. The City of Vancouver estimates that the average driver can meet their daily driving needs in under four hours using a Level 1 charger and in about 45 minutes using a Level 2 charger (new window) . Either of those is plenty of time if people have a place to park and charge overnight. Level 1 can even work for drivers in rural communities. Rob van Adrichem lives in Prince George, B.C., and got an electric car this past summer. He only has Level 1 charging at home, but tops up at Level 2 chargers at the park or the library in town if he needs to. I'm finding Level 1 is no problem, he said. I think people get scared off on Level 2s because they think it's going to be thousands of dollars and I don't know that it's always necessary. Is it a tenant's right to charge an EV at their rental? (new window) Ali Mohazab is co-founder of a startup called Parkizio Technologies that helps people such as apartment dwellers access electricity for charging. He said people thinking about switching to an EV may imagine doing a variation of what they did with their gas car: driving to empty and then going to a gas station and filling the entire tank — something they're forced to do because they don't have a gas pump at their home. Mohazab said that gas mentality may not allow people to see that with an EV, every parking opportunity is a charging opportunity and it doesn't matter if you charge faster so your battery is full at 1 a.m. instead of 8 a.m. when you leave for work — you can just leave it plugged in overnight. He added, If you kind of look at your car as a, you know, cell phone with wheels, then it really makes sense. McNally has found that she doesn't even need to charge every day, even at Level 1. I charge about once a week, she said. Couldn't be easier. But how can you tell if Level 1 will be enough for you? Level 1 is probably enough for most people, Mohazab says, except those who drive all day for work, such as Uber drivers. McNally suggests this rule of thumb: If you drive less than 60 kilometres a day, you are probably just fine with the regular plug that is already at your house. She recommends that new EV owners start with Level 1 to keep things cheap and simple. Start there, see how it works, learn your patterns — and then if you really want Level 2, you can add that cost later. Emily Chung (new window) · CBC News
National Post
03-06-2025
- National Post
H Company Launches Next-Generation Autonomous AI Agents for Enterprise and Consumer Markets
Article content Article content Article content H Company launches its next-generation European agent suite, comprising Runner H, Tester H, and Surfer H, and the open-sourcing of Holo-1 H, its cutting-edge visual-language model, lays the foundation for the next era of autonomous intelligence and artificial superintelligence. Three products reflect H Company's vision to make a trusted, action-oriented AI that delivers task execution beyond traditional chatbots, achieving a 92.2% success rate, the state-of-the-art in computer-use technology while reducing costs by up to 5.5x against peers (OpenAI, Anthropic, Google). The three AI agents cover task execution, web navigation and quality assurance, delivering faster task completion, GDPR-first data handling, and seamless hand-offs across the H-platform. Article content PARIS — H Company, a leading European AI Research and Product company backed by $220 million from tier-one investors, today announced the public launch of its autonomous AI agent platform featuring three specialized agents built to carry out complex digital tasks with human-in-the-loop. Article content The launch includes: Article content Runner H™, an orchestration platform for multi-step task completion, Surfer H™, a web-native agent for browser-based automation, Tester H™, an enterprise solution for automated software testing. Article content The company simultaneously open-sourced Holo-1, its proprietary visual-language model, in partnership with Hugging Face, Nvidia (NIM soon available) and AWS (Amazon). Article content H Company's current product selection supports its aim of making trusted, action-oriented AI agents available to organizations of all sizes. H Company's agents, which are built on proprietary multimodal models, bridge the gap by combining planning and vision capabilities to deliver measurable business outcomes. Article content 'We are witnessing the transition from conversational AI to execution intelligence,' said Charles Kantor, Co-founder and Chairman of H Company. 'Entering the era, where agents don't just assist—they complete entire workflows autonomously, delivering the productivity gains enterprises have been seeking from AI investments.' Article content Runner H serves as H Company's flagship orchestration platform, delivering 'intelligence in motion' by enabling users to accomplish complex, multi-step objectives through a single natural language prompt and is now available to the public. Article content Runner H dynamically coordinates specialized sub-agents and integrates with existing enterprise applications (including MCPs and agnostic interactions proprietary and additional agents), supporting seamless workflows across web browsers, document systems, and business applications while maintaining cost-effectiveness at scale. Article content Surfer H Article content Surfer H operates as a web-native automation agent, navigating browser environments with human-like precision while delivering cost-effective automation solutions. Unlike API-dependent solutions, Surfer H interacts directly with web interfaces, enabling deployment across any website without custom integrations or compatibility requirements, creating a powerful, general-purpose, cost-effective web automation system. Article content Tester H Article content Tester H addresses enterprise software testing challenges by converting natural language test scenarios into executable automated testing protocols. The platform supports both web and mobile applications, enabling teams to scale quality assurance processes without expanding technical testing resources. Article content Proven Performance Metrics Article content Benchmark testing demonstrates H Company's agents deliver superior cost-performance ratios and optimal value propositions compared to existing market solutions: Article content Surfer H achieved 92.2% task completion accuracy on the WebVoyager benchmark at just $0.13 per task, establishing the strongest tradeoff between accuracy and cost. Performance exceeds that of major technology providers while reducing operational costs by up to 10 times, setting the Pareto frontier for performance-per-dollar. Holo-1 visual-language model achieves 76.2% UI localization accuracy with affordable inference costs, leading performance among models under 7B parameters (the highest small-size model overall). Article content H Company's release of Holo-1 under Apache 2.0 licensing enables developers and enterprises to build production-ready custom automation solutions using visual AI capabilities. The model delivers commercial-grade accuracy for UI understanding and interaction. Article content The open-source initiative includes the WebClick dataset, featuring 1,639 real-world UI interaction scenarios designed to advance industry-wide development of visual AI applications. Article content Market Availability and Enterprise Adoption Article content Runner H is immediately available for public access, in waves, while Tester H targets enterprise customers requiring automated software testing capabilities. The platform's browser-based agent, Surfer H, ensures compatibility with existing enterprise privacy requirements, addressing key concerns for large-scale organizational deployment. Article content H Company is a next-generation AI research and product company pioneering the future of autonomous, agentic AI. Founded to build intelligence that acts—not just reacts—H Company is creating the foundational infrastructure for autonomous AI systems that drive real-world outcomes across industries. With $220 million in funding, H Company is backed by leading global investors including Accel, Eric Schmidt and Yuri Milner. Article content Article content Article content Article content Contacts Article content Article content Article content
