'What if Superintelligent AI Goes Rogue?' Why We Need a New Approach to AI Safety
As the world competes to create a true superintelligence, the United States government has begun removing previously implemented guardrails and regulation. The National Institute of Standards and Technology sent updated orders to the U.S. Artificial Intelligence Safety Institute (AISI). They state to remove any mention of the phrases "AI safety," "responsible AI," and "AI fairness." In the wake of this change, Google's Gemini 2.5 Flash AI model increased in its likelihood to generate text that violates its safety guidelines in the areas of "text-to-text safety" and "image-to-text safety."
If Superintelligence Goes Rouge
We are nearing the Turing horizon, where machines can think and surpass human intelligence. Think about that for a moment, machines outsmarting and being cleverer than humans. We must consider all worst-case scenarios so we can plan and prepare to prevent that from ever occurring. If we leave superintelligence to its own devices, Stephen Hawking's prediction of it being the final invention of man could come true.
AI apps are pictured.
AI apps are pictured.
Getty Images
Imagine if any AI or superintelligence were to be coded and deployed with no moral guidelines. It would then act only in the interest of its end goal, no matter the damage it could do. Without these morals set and input by human engineers the AI would act with unmitigated biases.
If this AI was deployed with the purpose of maximizing profit on flights from London to New York, what would be the unintended consequences? Not selling tickets to anyone in a wheelchair? Only selling tickets to the people that weigh the least? Not selling to anyone that has food allergies or anxiety disorders? It would maximize profits without taking into account any other factors than who can pay the most, take up the least time in boarding and deplaning, and cause the least amount of fuel use.
Secondarily, what if we allow an AI superintelligence to be placed in charge of all government spending to maximize savings and cut expenses? Would it look to take spend away from people or entities that don't supply tax revenue? That could mean removing spending from public school meal programs for impoverished children, removing access to health care to people with developmental disabilities, or cutting Social Security payments to even the deficit. Guardrails and guidelines must be written and encoded by people to ensure no potential harm is done by AI.
A Modern Approach Is Needed for Modern Technology
The law is lagging behind technology globally. The European Union (EU) has ploughed ahead with the EU AI Act, which at a surface glance appears to be positive, but 90 percent of this iceberg lurks beneath the surface, potentially rife with danger. Its onerous regulations put every single EU company at a disadvantage globally with technological competitors. It offers little in the way of protections for marginalized groups and presents a lack of transparency in the fields of policing and immigration. Europe cannot continue on this path and expect to stay ahead of countries that are willing to win at any cost.
What needs to happen? AI needs to regulate AI. The inspection body cannot be humans. Using payment card industry (PCI) compliance as a model, there needs to be a global board of AI compliance that meets on a regular basis to discuss the most effective and safe ways AI is used and deployed. Those guidelines are then the basis for any company to have their software deemed AI Compliant (AIC).
The guidelines are written by humans, but enforced by AI itself. Humans need to write the configuration parameters for the AI program and the AI program itself needs to certify the technology meets all guidelines, or report back vulnerabilities and wait for a resubmission. Once all guidelines are met a technology will be passed as AIC. This technology cannot be spot checked like container ships coming to port—every single line of code must be examined. Humans cannot do this, AI must.
We are on the precipice of two equally possible futures. One is a world where bad actors globally are left to use AI as a rogue agent to destabilize the global economy and rig the world to their advantage. The other is one where commonsense compliance is demanded of any company wanting to sell technology by a global body of humans using AI as the tool to monitor and inspect all tech. This levels the field globally and ensures that those who win are those that are smartest, most ethical, and the ones that deserve to get ahead.
Chetan Dube is an AI pioneer and founder and CEO of Quant.
The views expressed in this article are the writer's own.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Newsweek
6 hours ago
- Newsweek
Donald Trump Scores Major Legal Win in Accessing Sensitive Data of Millions
Based on facts, either observed and verified firsthand by the reporter, or reported and verified from knowledgeable sources. Newsweek AI is in beta. Translations may contain inaccuracies—please refer to the original content. A divided federal appeals court has ruled that the Trump administration's Department of Government Efficiency (DOGE) can access sensitive personal data held by several federal agencies, rejecting claims that the move violates privacy protections. In a 2-1 decision issued on Tuesday, the U.S. Court of Appeals for the Fourth Circuit vacated a lower court's preliminary injunction that had blocked DOGE-related personnel from obtaining administrator-level access to information at the Treasury Department, the Office of Personnel Management and the Department of Education. The ruling remands the case for further proceedings. Newsweek has contacted DOGE for comment via email outside regular office hours. The Justice Department declined to comment following the ruling. Why It Matters A divided federal appeals court ruling in favor of DOGE could significantly shift the balance between privacy protections and executive authority. The decision to grant embedded cross-agency teams broad, administrator-level access to sensitive personal data—including Social Security numbers, tax records and health information—strengthens the president's ability to direct internal modernization efforts across the federal bureaucracy. It also sets a potential precedent that could make it harder for unions, advocacy groups and individuals to challenge similar data-access policies in the future, narrowing judicial oversight when efficiency initiatives intersect with privacy concerns. Anti-DOGE protesters outside the Theodore Roosevelt Federal Building headquarters of the U.S. Office of Personnel Management in Washington, D.C., on February 5. Anti-DOGE protesters outside the Theodore Roosevelt Federal Building headquarters of the U.S. Office of Personnel Management in Washington, D.C., on February 5. Alex Wong/Getty What To Know The case stems from an executive order that President Donald Trump signed on January 20, which created DOGE to modernize "Federal technology and software to maximize governmental efficiency and productivity." The order directed agency heads to establish internal DOGE teams and provide "full and prompt access" to unclassified systems and records. Initially headed by Elon Musk, DOGE has been a controversial element of Trump's second term, overseeing spending and staffing cuts across agencies and facing multiple lawsuits. As a special government employee, Musk could serve in the role for only 130 days, and his tenure as the head of DOGE ended in May shortly before a public disagreement with the president. In February, U.S. District Judge Deborah Boardman granted a temporary restraining order and later a preliminary injunction limiting DOGE affiliates' access to certain data. The appeals court stayed that injunction in April pending appeal. The plaintiffs—a coalition that includes the American Federation of Teachers, several other labor unions and individual recipients of government benefits—had argued that granting DOGE affiliates such access violated the federal Privacy Act and the Administrative Procedure Act (APA). They said the data involved could include Social Security numbers, citizenship information, tax records, health histories and other personal identifiers. Judge Julius Richardson, joined by Judge G. Steven Agee, concluded that the plaintiffs had not shown a likelihood of success on the merits sufficient to justify preliminary relief. Writing for the majority, Richardson said, "The Privacy Act does not prohibit sharing information with those whose jobs give them good reason to access it." He also compared DOGE's broad modernization mandate to that of a consultant who must first survey systems to determine necessary improvements. The opinion also questioned whether the plaintiffs had standing to sue, noting that they had not alleged their specific records had been accessed, and whether the actions at issue constituted "final agency action" under the APA. The court further observed that the Privacy Act's civil remedies might preclude APA-based claims for equitable relief. Richardson's opinion cited a June U.S. Supreme Court order in a separate case that allowed DOGE access to Social Security Administration data while litigation continued. "This case and that one are exceedingly similar," Richardson wrote, adding that the precedent informed the court's equitable discretion. In dissent, Judge Robert King argued that the district court acted "quickly—but extremely carefully" in blocking DOGE's access given the scope and sensitivity of the data. King warned that the executive order had granted "unfettered, unprecedented, and apparently unnecessary access" to personal information for millions of Americans and criticized the majority for adopting what he described as a "heightened standard" for likelihood of success. The unions involved, which include the National Federation of Federal Employees and the International Association of Machinists and Aerospace Workers, did not immediately respond to requests for comment. What People Are Saying Aman George, senior counsel at Democracy Forward, commenting on a different federal court ruling that declined to block DOGE's access to health and labor data, said in a news release on June 27: "While today's decision is disappointing, the court made clear it shares our deep concerns. We are committed to continuing this case and holding the administration accountable for exposing millions of Americans' private records to politically motivated operatives with no legal authority to access them." Justice Ketanji Brown Jackson, joined by Justice Sonia Sotomayor, sharply criticized the Supreme Court's June decision, writing in dissent: "In essence, the 'urgency' underlying the government's stay application is the mere fact that it cannot be bothered to wait for the litigation process to play out before proceeding as it wishes." What Happens Next The appeals court's decision does not end the litigation. The case returns to the district court for further proceedings on the plaintiffs' underlying claims. The outcome represents a significant legal victory for the administration's DOGE initiative, reinforcing earlier high-court signals that agency-embedded DOGE teams may access certain records to perform modernization work. Still, the broader legal battle over the scope of such access—and its compatibility with privacy protections—remains unresolved.
Tom's Guide
a day ago
- Tom's Guide
National Public Data is giving out your address and phone number — here's how to stop them
Though it has been dormant for most of the year, the website National Public Data has reemerged with new owners. As reported by PCMag, the website was well known for a major data breach that occurred last year in which millions of Social Security numbers were leaked online, resulting in a barrage of lawsuits against the former site owners, Jericho Pictures. The breach exposed at least 272 million Social Security numbers and 600 million phone numbers and after the lawsuits were filed, the site went dormant until recently when it went live again as a people finder site. Although there is nearly no information about the new owners of the domain aside from it being registered to a Florida-based VPN service called 'Perfect Privacy,' it's acting as a free search engine to let people search for information on others. The new National Public Data lets users look up anyone's personal information including addresses, phone numbers, birthdates, relatives, location, age, workplace and criminal records – all for free. While the new site claims to have no affiliation with the previous owners, they're also not very forthcoming about where they're gathering their data aside from stating that it's being collected 'from publicly available sources including federal, state and local government agencies, social media pages, property ownership databases and other reliable platforms. After the data is in our hands, we verify and filter it to make sure it is indeed accurate and up-to-date.' Now, just because they may be getting the data from legitimate sources and verifying it doesn't mean that you want them to have it or be make it readily accessible and easy to find online. If you would like National Public Data to remove your data from its site, you have that option as there is an opt out form you can use to delete your profile data. Get instant access to breaking news, the hottest reviews, great deals and helpful tips. When PCMag tried it, they reported that it seemed as though their data was immediately removed from the site. However, it does appear that many users would likely be unaware of this option, which makes it easy for sites like this to continue posting their information online. Additionally. there are certainly data removal services like Incogni that can help you delete your online data as well as identity monitoring services that can help you monitor where you data appears online. 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.
Yahoo
3 days ago
- Yahoo
I Asked ChatGPT If Universal Basic Income Could Replace Social Security: Here's What It Said
It's a straightforward thought, and, at first glance, it seems reasonable: replace Social Security with a universal basic income (UBI). It not only could help lower-income retirees live more comfortably in their later years and erase retirement fears, but also could lift lower-income Americans out of poverty for life. Read Next: Check Out: But would it work? GOBankingRates asked ChatGPT if universal basic income could replace Social Security, and here's what it said. First Thoughts From ChatGPT When asked the straightforward question, with no background in the prompt, 'Could universal basic income ever replace Social Security?' ChatGPT offered a simple, well-organized answer. 'That's a big, fascinating question — and the short answer is: not easily, and not anytime soon,' it said. However, it added, 'UBI isn't a miracle cure or a looming catastrophe — it's a tool with real potential, if designed carefully and funded wisely.' It went on to explain why. Explore More: Social Security vs. UBI, According to ChatGPT First, ChatGPT succinctly explained the differences between the programs. 'Social Security is a targeted social insurance program, funded by payroll taxes, that provides retirement, disability and survivor benefits,' it said. On the other hand, UBI is a universal cash transfer, where 'everyone gets the same amount, regardless of income or work history.' Replacing Social Security with UBI would require a paradigm shift. 'Replacing Social Security with UBI would mean shifting from an earned benefit model to a universal entitlement, which is a huge philosophical and political leap,' ChatGPT said. Expert Thoughts Forbes writer Cornelia Walther, Ph.D., an AI researcher, explored how technology could assist in UBI distribution. Yet, she acknowledged in a recent article that UBI acceptance requires 'a foundational human shift.' 'It requires a societal commitment to move beyond paradigms of scarcity and competition,' she wrote. Will People Stop Working With UBI? ChatGPT was then asked this question: 'If there is a universal basic income, will people want to work less?' ChatGPT responded that there have been many studies done with UBI pilots and similar programs, which largely showed that 'most people don't stop working, but some do adjust how they work.' ChatGPT cited several real-world examples from Finland, Canada, California and Alaska, with the takeaway, 'The evidence shows reductions are modest and often socially beneficial.' Expert Thoughts The world's largest study to date, funded partially by Sam Altman, conducted by OpenResearch and reported by the National Bureau of Economic Research, contradicted previous findings. In the study, 1,000 low-income individuals received $1,000 per month for three years. A control group of 2,000 people received $50 per study found that labor market participation decreased by 2 percentage points and participants worked roughly 1.3 to 1.4 fewer hours per week. Partners of participants also reduced their work time similarly. Participants actually earned $1,500 less each year, after accounting for the UBI deposits. Unemployment lasted 1.1 months longer for those receiving the $1,000 monthly check, which seemed to contradict prior research. Nevertheless, the bulk of research, as ChatGPT accurately cited, showed increases in life satisfaction and overall health. The Costs ChatGPT dug into the numbers to address the next challenge: costs and how to pay for UBI. 'Social Security is expensive already, but UBI would dwarf its cost,' it said. Citing figures from the Social Security Administration, it said that Social Security outlays for Supplemental Security Income and Social Security Disability Insurance would total roughly $1.5 trillion in 2025. 'Unless UBI replaced all major welfare programs, the U.S. would need massive new taxes or deficit spending,' ChatGPT said. Expert Thoughts Teddy Ganea, a writer at The Stanford Review, shared that it's entirely possible to implement a UBI of $18,000 per year, at minimum, that could 'end poverty overnight.' The program would phase out gradually for higher-income individuals. Ganea explained that a UBI that provides roughly $9,000 on average in assistance to the 75% of Americans who made less than $75,000 in 2024 would cost less than the $2.5 trillion currently spent on programs like Medicaid and welfare. The cost would even leave enough to 'bolster underfunded programs like Social Security or education,' Ganea wrote. Bottom Line In the ultimate irony, just as ChatGPT shared reasons UBI is impractical, Ganea pointed to the same generative AI as a catalyst for UBI. 'Maybe ChatGPT's greatest achievement won't be in automating coding or customer service,' Ganea wrote. 'Perhaps it will be that, in stoking fears of automation, it paved the way for UBI.' More From GOBankingRates New Law Could Make Electricity Bills Skyrocket in These 4 States I'm an Economist: Here's When Tariff Price Hikes Will Start Hitting Your Wallet 5 Strategies High-Net-Worth Families Use To Build Generational Wealth How Far $750K Plus Social Security Goes in Retirement in Every US Region This article originally appeared on I Asked ChatGPT If Universal Basic Income Could Replace Social Security: Here's What It Said Sign in to access your portfolio
