Future of AI According to Daniela Amodei of Anthropic

Some may argue that regulations stifle innovation in technology, but a key figure in the AI industry believes otherwise.

During a conversation at WIRED’s Big Interview event, Daniela Amodei, President and Co-founder of Anthropic, discussed with editor at large Steven Levy how her company is navigating the Trump administration's policies. Although some speculate that Anthropic’s emphasis on AI risks could be interpreted as fearmongering aimed at regulatory manipulation, Amodei stands firm in her belief that highlighting AI dangers strengthens the field.

Amodei emphasized, “From the beginning, we were vocal about the vast promising horizon of AI. We want global communities to experience its potential and advantages, which requires addressing the challenging aspects and keeping risks under control. Hence, our continuous dialogue on these points.”

Anthropic’s Claude model boasts over 300,000 users from startups and corporations, reflecting how they engage with diverse clients. Amodei has found that users not only desire powerful AI capabilities but also value dependability and safety.

“Nobody expresses a desire for less safety in a product,” Amodei stated, comparing Anthropic’s transparency about its model’s limitations to car manufacturers sharing crash-test outcomes. It's similar to how a safety improvement following a crash test can assure potential buyers. This transparency enables a market that partly self-regulates, akin to the auto industry.

“We are inadvertently setting baseline safety benchmarks through our economic contributions,” she explained. “[Corporations] are integrating AI into many daily operations, recognizing the consistency of our product in minimizing hallucinations and avoiding harmful outputs, unlike competitors with less emphasis on safety.”

Anthropic is notable for its 'Constitutional AI' philosophy, wherein models are trained using core ethical guidelines and human values. By employing principles like those from the United Nations Universal Declaration of Human Rights, a language model can craft responses that align more with overarching ethical perspectives than mere factual correctness.

This ethical dedication not only informs their product but aids in retaining skilled personnel, Amodei noted. “Individuals joining us share stories about resonating with our mission and values, appreciating our honesty in both successes and challenges, and the desire to improve negative aspects genuinely.”

This ethos likely explains why Anthropic’s workforce has swelled from 200 to over 2,000 employees. Despite talks of an AI bubble looming over finance and tech sectors, Amodei reports no slowdown in company growth.

“The progress we see in model sophistication aligns with scaling predictions, and likewise, our financial performance,” Amodei remarked. “Our scientists understand the trajectory continues until it doesn’t, so we maintain an attitude of humility and self-awareness.”