Should AI Be Banned? The Case for and Against a Ban

Should AI Be Banned? Where the Argument Stands

No country bans AI outright. The live debate is over banning specific uses of AI and over banning superintelligence, meaning AI capable of surpassing human cognition.

Targeted bans already exist. The European Union, several US cities, Illinois and New York City have each written one into law or policy.

If you think AI should be stopped or limited, the bans below show which uses governments and school systems have already stopped.

Where Specific Uses of AI Are Already Banned

Existing AI bans each cover one use, one product or one place, such as social scoring in the EU or ChatGPT in Italy for four weeks.

The European Commission says the prohibited practices "entered into application from 2 February 2025." The eight practices are:

• Harmful AI-based manipulation and deception.
• Harmful AI-based exploitation of vulnerabilities.
• Social scoring.
• Individual criminal offence risk assessment or prediction.
• Untargeted scraping of the internet or CCTV material to create or expand facial recognition databases.
• Emotion recognition in workplaces and education institutions.
• Biometric categorisation to deduce certain protected characteristics.
• Real-time remote biometric identification for law enforcement in publicly accessible spaces.

Article 5 allows real-time police face recognition only where it is "strictly necessary" for three purposes.

• Finding victims of abduction, trafficking or sexual exploitation, or missing persons.
• Preventing a specific, substantial, imminent threat to life or a terrorist threat.
• Locating a suspect of a listed serious crime punishable by at least four years.

Use needs prior judicial or independent authorization, with a 24-hour exception for urgent cases. Article 5 does not cover private or after-the-fact face recognition. More detail is in our guide to the EU AI Act.

A ninth ban, added by the AI Omnibus, covers AI that generates nonconsensual sexual content and child sexual abuse material. The Commission says it comes into effect in December 2026.

US city bans on facial recognition began with San Francisco in May 2019. San Francisco's Board of Supervisors voted 8 to 1 in May 2019 to bar city agencies and police from using facial recognition. The ordinance does not regulate private businesses.

Boston's City Council voted unanimously on June 24, 2020 to ban city use of facial recognition. Portland, Oregon followed on September 9, 2020. Its vote also restricted private businesses from using facial recognition in public areas, a rule that took effect January 1, 2021.

Illinois prohibits anyone from using AI to make mental health and therapeutic decisions. The state allows AI for "administrative and supplementary support services" for licensed professionals.

IDFPR Secretary Mario Treto Jr. said that Illinois people "deserve quality healthcare from real, qualified professionals and not computer programs."

Public Citizen counted 25 states with election deepfake laws by May 2025, which it called half of all states, up from five before 2024. Its tracker, updated October 5, 2026, lists enacted laws in 33 states. Public Citizen doesn't give a total, so 33 is our count of the tracker's entries. Many of these laws regulate deepfakes, for example by requiring disclosure, and do not ban them outright.

In July 2025, the Senate voted 99 to 1 to strip a proposed moratorium on state AI laws from the One Big Beautiful Bill Act. Goodwin reported that the moratorium would have halted more than 1,000 pending state AI bills.

China is a separate case. Its regulators issued interim measures for generative AI services in July 2023, effective August 15, 2023. Morrison Foerster describes the rules as regulation that also aims to foster AI development.

OpenAI's supported-countries list does not include China, Iran, North Korea or Russia. The Register reported that OpenAI blocks API traffic from regions off that list. In those four countries, OpenAI chose not to offer its service. No government ban on AI is involved in those cases.

The Case for Banning or Pausing AI

The Future of Life Institute's 2023 open letter asks for a pause, and the 2025 Statement on Superintelligence asks for a prohibition.

The Future of Life Institute (FLI) published an open letter dated March 22, 2023. It says: "We call on all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4."

The letter's page showed 31,810 signatures as of October 14, 2026.

The Statement on Superintelligence came later, on October 22, 2025. It had nearly 65,000 signatures at release. Its site showed 77,550 on October 14, 2026, including 5,000 from an Ekō petition.

The statement says the prohibition stays in place until scientists broadly agree superintelligence can be built safely and controllably. It also requires strong public buy-in.

Signers include Geoffrey Hinton, Yoshua Bengio, Stuart Russell, Steve Wozniak, Richard Branson, Steve Bannon and Glenn Beck. FLI's October 2025 newsletter also lists researchers from OpenAI, Anthropic, Google DeepMind and Meta among the signers.

On September 3, 2026, Sen. Bernie Sanders and Rep. Greg Casar announced a proposed Ban Artificial Superintelligence Act. Crypto Briefing reported that the bill had not been formally introduced as of early September. No bill number has been confirmed.

The bill would ban superintelligent AI capable of surpassing human cognition and resisting shutdown. It would also pause advanced AI until a new federal regulator sets safety rules.

Data for Progress polled more than 1,300 likely voters on the proposal. Common Dreams reported on September 10, 2026 that 68 percent supported it and 25 percent opposed. Support was 72 percent among Democrats, 70 percent among independents and 63 percent among Republicans.

PauseAI was founded in May 2023 in Utrecht by Joep Meindertsma. StopAI was founded in 2024. Our guide to PauseAI and StopAI covers both groups.

In April 2023, YouGov asked 20,810 US adults whether they would support or oppose a six-month pause on some kinds of AI development. Of them, 41 percent strongly supported it and 28 percent somewhat supported it.

Nine percent somewhat opposed the pause, 4 percent strongly opposed it and 18 percent were not sure. The question came right after a mention of tech leaders' open letter, so read YouGov's full wording before relying on the result.

FLI wrote the statement and ran its own US poll, released with it. The poll found 64 percent said superintelligence should not be developed until it is shown to be safe and controllable. It found 73 percent wanted robust regulation of advanced AI and 5 percent backed the status quo of unregulated development.

The Case Against a Ban

Opponents of a wider ban, including JD Vance and Andrew Ng, argue that regulation slows useful work and that extinction warnings help established companies.

At the Paris AI Action Summit, JD Vance argued that AI's promise will never come to pass if overregulation deters innovators.

The United States did not sign the summit declaration, which more than 60 nations signed, including China. The United Kingdom also declined to sign.

Andrew Ng argued in October 2023 that large tech companies are creating fear of AI leading to human extinction to avoid competing with open-source AI. In his view, extinction framing drives regulation that protects incumbents.

AI also has medical uses. The FDA's list of AI-enabled medical devices has more than 1,600 entries. The list's most recent final decision is dated June 29, 2026.

The FDA's list is current as of September 4, 2026.

Some restrictions in the table were short-lived. Italy's ChatGPT block lasted about four weeks, and NYC's 2023 ChatGPT block was walked back within about five months. Those cases do not show that every ban fails.

Should AI Be Banned in Schools?

New York City shows what a school restriction looks like in practice. The city blocked ChatGPT in January 2023, then limited student-facing generative AI in 2026.

A spokesperson cited "concerns about negative impacts on student learning, and concerns regarding the safety and accuracy of content."

In May 2023, Chancellor David Banks wrote that "the knee-jerk fear and risk overlooked the potential of generative AI to support students and teachers." He described a shift to exploration and careful examination.

On September 2, 2026, Mayor Zohran Mamdani and Chancellor Kamar Samuels announced a one-year moratorium. It covers student-facing generative AI from 2-K through eighth grade in the 2026 to 2027 school year.

The moratorium covers nearly 600,000 students, about two-thirds of the system. Companion chatbots are banned in all grades.

High schoolers get five supervised pilots, capped at 45 minutes a week. General chatbots such as ChatGPT and Claude are excluded from those pilots.

Exceptions cover assistive technology in IEP and 504 plans, a DOE-approved translation app, and career programs such as computer science. About 40 previously allowed programs with AI features will have those features disabled or removed.

NYC chose a limit by grade and by use. Our guide to whether AI is banned in schools says no state has a total ban on generative AI in public schools, and there is no federal prohibition.

What You Can Do If You Want Limits

Every ban in the table names what it covers and when it began. A limit you propose should name both, so anyone can see what it stops and from when.

1. Pick one use, such as police face recognition, mental health decisions or student chatbots, instead of AI as a whole.
2. Read the rule that already covers that use. The text of Article 5 is the EU rule, and the Illinois and New York City rules above show US examples.
3. Match the level of government to the use. The table shows city, state, EU and school rules. At the federal level, the Sanders and Casar bill is a proposal whose formal introduction is unconfirmed.

For how AI rules are set, see our guide to AI regulation. For ways to push back, see our guide to fighting back against AI. For switching off AI features in the products you use, see our guide on how to turn off AI.

Sources

• AI Act, Article 5 (consolidated text), artificialintelligenceact.eu
• AI Act regulatory framework, European Commission
• San Francisco becomes first city in US to ban facial recognition software, ScienceAlert, May 2019
• Boston bans facial recognition due to concern about racial bias, VentureBeat, June 24, 2020
• Portland passes nation's toughest restriction on facial recognition technology, OPB, September 9, 2020
• Gov. Pritzker Signs State Legislation Prohibiting AI Therapy in Illinois, Illinois IDFPR, August 4, 2025
• 25 States Enact Laws to Regulate Election Deepfakes, Public Citizen, May 13, 2025
• Tracker: State Legislation on Deepfakes in Elections, Public Citizen, updated October 5, 2026
• NYC schools ban ChatGPT, Chalkbeat, January 3, 2023
• David Banks on ChatGPT, Chalkbeat, May 18, 2023
• Mamdani AI moratorium in NYC public schools, amNY, September 2026
• Pause Giant AI Experiments: An Open Letter, Future of Life Institute
• Statement on Superintelligence
• FLI newsletter, October 2025
• Fortune on the superintelligence statement, October 22, 2025
• YouGov daily results on a six-month AI pause, April 3, 2023
• Common Dreams on the Sanders and Casar bill poll, September 10, 2026
• Ban Artificial Superintelligence Act, Crypto Briefing, September 3, 2026
• ChatGPT available to users in Italy a month after temporary ban, Al Jazeera, April 28, 2023
• OpenAI blocks API traffic from unsupported countries, The Register, June 25, 2024
• China: Interim Measures Governing Generative AI Issued, Morrison Foerster, July 2023
• Vance rails against excessive regulation at Paris AI summit, PBS NewsHour, February 11, 2025
• Federal AI Moratorium Dies on the Vine, Goodwin, July 2025
• List of AI-Enabled Medical Devices, FDA
• Google Brain founder Andrew Ng on extinction fears, SiliconANGLE, October 31, 2023