ChatGPT Agent supercharges AI to carry out tasks — here's how OpenAI's new agent works
With the agent mode activated, users can ask ChatGPT to perform tasks such as:
Unlike typical chatbot interactions that involve generating text, the agent can navigate websites, securely log in with user permission, run code, compile research into spreadsheets or slides and deliver outputs in editable formats.
Users retain control throughout the process, with ChatGPT requesting explicit approval before submitting forms or handling sensitive information.
The new capability builds on OpenAI's previous tools by combining:
Together, these components enable the AI to transition between reasoning and action, completing tasks in a more autonomous and organized way while still prompting users when clarification or approval is needed.
To enable the feature, users can open any ChatGPT conversation and select 'agent mode' from the tools dropdown.
Get instant access to breaking news, the hottest reviews, great deals and helpful tips.
Once active, the system can carry out multi-step workflows that typically require switching between apps, browser tabs or tools.
OpenAI has emphasized the safety systems built into this release. The ChatGPT agent is designed to avoid high-risk actions, such as sending emails, making purchases or offering legal or financial advice, without user approval.
It has been trained to recognize and reject malicious or ambiguous instructions, and it alerts users to any uncertainty or potentially sensitive actions.
To further reduce risks, OpenAI has implemented a range of safeguards, including always-on classifiers, refusal training for dual-use scenarios and enforcement pipelines to prevent misuse, particularly in areas involving biological or chemical threats.
These protections reflect OpenAI's broader Preparedness Framework, which prioritizes caution even in the absence of direct evidence that the model could enable harmful activity.
'We don't have direct evidence the model could help a novice create severe biological or chemical harm,' OpenAI noted, 'but we are exercising caution.'
According to OpenAI, this release is an early step in expanding agentic AI capabilities. The company plans to regularly add new features and improvements over time, with the goal of making ChatGPT more versatile and useful for a broader set of users.
Whether for summarizing meetings, conducting research or preparing presentations, ChatGPT's new agent functionality signals a shift from conversational assistance to hands-on task execution. Fortunately, oversight and control remain in the user's hands.
Follow Tom's Guide on Google News to get our up-to-date news, how-tos, and reviews in your feeds. Make sure to click the Follow button.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
17 minutes ago
- Yahoo
OpenAI launches $50 million fund to support nonprofits, community organizations
By Anna Tong SAN FRANCISCO (Reuters) -ChatGPT maker OpenAI is launching a $50 million fund to support nonprofit and community organizations, the artificial intelligence company said on Friday. The fund is the first action following a recommendations report from the San Francisco-based company's nonprofit commission, which was formed in April to guide OpenAI's philanthropic efforts. OpenAI has been working to revamp its corporate structure, which it says is necessary in order to continue raising the massive amounts of capital needed to stay competitive in the AI arms race, a move it is trying to balance with its founding mission, as a nonprofit, to develop AI for the public good. Currently, its nonprofit arm owns and controls its for-profit arm, and OpenAI plans to convert the for-profit entity into a public benefit corporation, of which the nonprofit parent would become a shareholder. The nonprofit commission was established as part of OpenAI's efforts to show that it can remain true to its founding mission despite the corporate revamp. OpenAI said the fund will facilitate partnerships to implement AI in sectors such as education, economic opportunity, community organizing and healthcare. It will also back community-led research and innovation focused on using AI for public good. The nonprofit commission submitted its recommendations report Thursday, following interviews with over 500 nonprofits and community experts.
Business Insider
19 minutes ago
- Business Insider
JPMorgan pushes into private company coverage as Wall Street eyes growing opportunity
"This new offering aims to expand our coverage and research into private companies, where we continue to see growing influence and interest from our clients," Hussein Malik, the bank's head of global research, said in an internal memo. "Importantly, private companies are becoming increasingly relevant to various industries, especially in the new economy space," the memo said. The move comes as more company founders and boards choose to stay private longer. The median age of a private company going public has increased from 6.9 years a decade ago to 10.7 years today, according to Morningstar. A major factor behind the decline is the surge in funding for companies like OpenAI and SpaceX, fueled by the rise of private equity and venture capital. Pitchbook estimates that there are $18.7 trillion assets in private markets, including venture capital and private equity — a figure that will reach $24 trillion by the end of 2029. "With approximately 1,200 private companies in the US achieving unicorn status between 2020 and 2023—a notable increase from around 330 between 2016 and 2019—their growing influence on the economy and markets is clear," Malik said in a separate memo to clients. Malik said understanding the private markets has now become key to properly assessing publicly traded companies. "Understanding their impact is and will remain crucial for both public and private market investors to make informed investment decisions," he added. Jamie Dimon has long lamented what he has called "the diminishing role of publicly traded companies in the American financial system." In a 2023 letter to investors, he said the number of US public companies had declined to 4,300 from a peak of 7,300 in 1996.
Forbes
36 minutes ago
- Forbes
From Tools To Teammates: How AI Agents Will Become Digital Labor
OpenAI just released its ChatGPT Agent - is this the beginning of AI agents going mainstream? The future of work just arrived. On July 17, 2025, OpenAI launched ChatGPT Agent, marking a pivotal moment in artificial intelligence evolution. This isn't just another AI chatbot. This is the beginning of digital labor. Think of having a digital colleague that can now create presentations, navigate websites, conduct deep research, and complete complex tasks on their own (in AI speak, 'autonomously'). For businesses and consumers alike, this represents a fundamental shift in how everyone will work, shop, and interact with technology. Understanding agentic AI's game-changing potential The numbers tell a compelling story: Grand View Research estimates the global AI agents market is set to explode from $5 billion in 2024 to $50 billion by 2030, a 46% compound annual growth rate. More importantly, according to new research from the Capgemini Research Institute, AI agents could generate up to $450 billion in economic value by 2028 through revenue growth and cost savings. Yet despite this massive opportunity, only 2% of organizations have deployed AI agents at scale, creating a narrow window for competitive advantage that won't remain open for long. Unlike traditional AI that responds to prompts, agentic AI possesses genuine 'agency' - the ability to set goals, make decisions, and take actions with minimal human oversight. Harvard Business Review describes these systems as having "supercharged reasoning and execution capabilities" that go far beyond simple question-answering to actually performing complex tasks. The distinction is crucial: while generative AI is more about language to language and creates content, agentic AI is about multi-step reasoning, planning and it acts. It can book your flights, process insurance claims, manage inventory, and even conduct comprehensive research across hundreds of sources. This autonomous capability transforms AI from a tool into a true digital teammate. What Exactly Is an AI Agent? Unlike traditional AI that responds to prompts, an AI agent is artificial intelligence that handles multistep tasks without requiring a human to steer it the whole time. This is now the next phase of the AI era - 'Agentic AI'. While ChatGPT answers questions, AI agents actually do things - they book flights, process invoices, debug code, and conduct research across hundreds of sources autonomously. An example of OpenAI's ChatGPT Agent in action. The key differentiator: agents can take multiple actions, connect to various applications, and work for extended periods. OpenAI's Codex agent can work for up to 30 minutes without human supervision, while Anthropic's Claude 4 can tackle coding problems for up to seven hours straight. The Seven Species of Digital Workers While there will eventually be millions of agents, let's try to organize them into the distinct types of AI agents that are now entering the workforce. The Information had a nice way to summarize the different kinds of digital labor: What they do: Handle enterprise workflows across multiple software applications Digital labor: Invoice processing, data entry, document classification, scheduling Examples: UiPath, Microsoft Power Automate, Zapier + AI What they do: Resolve customer support and employee questions through dialogue Digital labor: Customer service, IT tickets, HR tasks Examples: Salesforce Agentforce, ServiceNow NowAssist, Sierra, Decagon What they do: Retrieve, analyze, and validate information from trusted sources Digital labor: Academic research, citation sourcing, technical analysisExamples: OpenAI Deep Research, Perplexity Pro, Scite Assistant, AlphaSense What they do: Analyze data to produce graphics, charts, and reports Digital labor: Data querying, dashboard creation, business insights Examples: Power BI Copilot, Tellius, ThoughtSpot, Glean What they do: Handle complex coding tasks for software engineers Digital labor: Code completion, debugging, documentation, site reliability Examples: Cursor, GitHub Copilot, Claude Code, Cognition's Devin What they do: Specialized work in regulated fields like law, medicine, finance Digital labor: Contract analysis, medical triage, financial analysis Examples: Harvey (legal), Hippocratic AI (healthcare), Rogo and Hebbia (finance) What they do: Navigate websites and handle repetitive online tasks Digital labor: Form filling, online ordering, social media posting Examples: OpenAI Operator, Google Project Mariner, Anthropic Computer Use OpenAI's bold vision becomes reality OpenAI's agent rollout began with Operator in January 2025, an AI capable of using web browsers like humans - clicking buttons, filling forms, and navigating websites. Then came Deep Research in February, which analyzes hundreds of sources to generate fully-cited reports in minutes. The July launch of ChatGPT Agent unified these capabilities, creating what The Wall Street Journal calls "an agent that can make spreadsheets and PowerPoints" while handling complex multi-step workflows. Sam Altman, OpenAI's CEO, predicts these agents will "materially change the output of companies" in 2025, estimating they can already handle "a single-digit percentage of all economically valuable tasks in the world." With 41.6% accuracy on complex reasoning benchmarks (double previous models) these agents represent a quantum leap in AI capability. Transforming experiences across consumer and business landscapes AI agents are revolutionizing both consumer experiences and business operations at unprecedented scale. For consumers, the transformation is happening at remarkable speed: recent reports show 95% of customer interactions are predicted to be handled by AI in 2025, while current deployments show AI-powered systems reducing resolution times by up to 52%and delivering 31.5% higher customer satisfaction scores compared to traditional support methods. The consumer impact extends far beyond convenience. Klarna's AI assistant reduced average customer issue resolution from 11 minutes to just 2 minutes while maintaining customer satisfaction scores equal to human agents. Virgin Money's AI assistant "Redi" has handled over 2 million customer interactions with a 94% satisfaction rate, demonstrating that consumers readily embrace AI-powered service when it delivers superior results. The retail sector shows equally impressive adoption, with 24% of consumers already comfortable with AI agents making purchases on their behalf—a figure that jumps to 32% among Gen Z shoppers, while 75% of customer inquiries can now be resolved by AI tools without human intervention. The business case for AI agents is equally compelling and backed by remarkable real-world results. Organizations implementing AI report 6-10% average revenue increases, with 62% of companies expecting full 100% or greater returns on investment. The operational improvements are staggering: companies report 83% experiencing revenue growth versus 66% without AI implementation, 76% improvement in operational efficiency, and financial institutions seeing increases in profitability through enhanced fraud detection and personalized service. Real-world success stories illustrate the transformative potential across industries. JPMorgan Chase's AI-driven "Coach" tool helps wealth advisers retrieve research 95% faster, contributing to a 20% year-over-year increase in asset management sales. The bank's AI initiatives have already saved nearly $1.5 billion through fraud prevention and operational efficiencies. Wiley achieved a 40% increase in case resolution with AI agents, while 76% of e-commerce teams credit AI with revenue growth and 92% of service teams report cost reductions. Manufacturing leaders report 40% reduction in downtime through AI-driven predictive maintenance. Employee productivity transformation is equally impressive, ranging from customer service agents answering more inquiries per hour, business professionals writing more documents per hour, and programmers coding more projects per week using AI agents. These are just early use cases, but you can already see how agentic ai will fundamentally redefine what exceptional customer experiences and business performance looks like. Why This Is Just the Beginning We're in the early innings of digital labor. Current agents still make mistakes and require human oversight, but they're evolving rapidly. The combination of cheaper reasoning models, better orchestration software, and expanding application integrations means agent capabilities are compounding quickly. The workforce of 2030 won't just include humans - it will be a hybrid ecosystem where digital agents handle routine tasks while humans focus on creativity, strategy, and relationship-building. We're not just automating work; we're creating a new category of digital colleague that augments human capability rather than simply replacing it. The age of digital labor has begun. The question isn't whether these AI agents will transform work - it's how quickly businesses and consumers will adapt to this new reality.
