Will Halt Potential Rogue AI Systems, Says Microsoft AI Chief
Mustafa Suleyman further noted that now the focus was on ‘humanist superintelligence,’ systems that are strictly designed to serve human interests rather than operate autonomously.
Topics
News
- Will Halt Potential Rogue AI Systems, Says Microsoft AI Chief
- Google Bets Big on Data Infra With a $4.75B Deal in Intersect
- OpenAI Softens ChatGPT’s Tone While Scaling for an AI Showdown
- UAE President Meets Musk to Discuss AI and Cross-Border Collaboration
- UAE Forms National Media Authority to Centralize Media Oversight
- TikTok Strikes US Sale Deal With Oracle, Silver Lake, and UAE’s MGX
After the expiration of its 18 month contract, which had Microsoft barred for independent artificial general intelligence (AGI) research and development, a new renegotiated definitive agreement with OpenAI will see Microsoft develop superintelligence individually or in collaboration with third parties.
Mustafa Suleyman, the AI chief at Microsoft, has shared that the tech giant will halt development of any AI system that has the potential of going rogue. “We won’t continue to develop a system that has the potential to run away from us,” he shared with Bloomberg, adding that containment and alignment were “necessary prerequisites” and “red lines” for releasing superintelligent tools.
Additionally, Microsoft recently formed a new superintelligence team under Suleyman. By aggressively pursuing AGI, systems capable of performing tasks at par with humans, Microsoft positions itself in a more direct competition with its collaborator OpenAI and competitors such as Anthropic, Google, and Meta.
Suleyman further noted that now the focus was on ‘humanist superintelligence,’ systems that are strictly designed to serve human interests rather than operate autonomously.
The AGI team will be focused on building a “world-class, frontier-grade research capability in-house,” Mustafa told Business Insider. “Microsoft needs to be self-sufficient in AI,” he said. “And to do that, we have to train frontier models of all scales with our own data and compute at the state-of-the-art level.”
He drew a comparison of Microsoft with that of a “modern construction company,” as thousands of workers build gigawatts of CPUs and AI accelerators.
On a lighter side, Suleyman finds chatbots as a good way for one to offload their emotions and detoxify themselves. Talking on Mayim Bialik’s “Breakdown” podcast, he pointed out that companionship and support have gained the status of AI’s most popular use cases.
”That’s not therapy,” Suleyman said. “But because these models were designed to be nonjudgmental, nondirectional, and with nonviolent communication as their primary method, which is to be even-handed, have reflective listening, to be empathetic, to be respectful, it turned out to be something that the world needs.”
On the other hand, he feels that “this is a way to spread kindness and love and to detoxify ourselves so that we can show up in the best way that we possibly can in the real world, with the humans that we love.”
Suleyman joined Microsoft in March 2024 after co-founding DeepMind and Inflection.
