Accenture partners with Anthropic to launch first embedded AI evaluator

Dario Amodei’s initiative to embed independent safety evaluators within AI laboratories is materializing: Anthropic has announced that employees from technology consulting firm Accenture will operate on-site to audit its models and personnel.
According to a blog post, Anthropic stated that Faculty, an AI-focused subsidiary acquired by Accenture in January, will commence “evaluating and red-teaming models, conducting alignment assessments, and testing model safeguards.” Both organizations anticipate investing a minimum of $1 billion into this initiative over the next five years.
Anthropic’s selection of Accenture caught many AI observers—and financial markets—by surprise, causing the consultant’s stock to surge 8% after hours. The conversation surrounding embedded evaluators, sparked by Amodei’s blog post, has largely centered on specialized AI safety research groups such as METR, Redwood Research, and Apollo Research. This focus is especially pronounced at Anthropic, which places AI safety and alignment at the core of its mission.
Anthropic indicated that additional evaluators will be announced in the coming weeks and confirmed it is in discussions with METR and other non-profit organizations regarding how to “pilot elements of embedded evaluation using their own funding.”
Although Accenture is not renowned for cutting-edge deep learning research, Anthropic highlighted the firm’s practical experience in deploying AI for major corporations and government agencies as a significant advantage. Furthermore, as a large public company that predates the current AI boom, it maintains greater functional independence from Anthropic and the complex ecosystem surrounding the AI laboratory.
The laboratory noted that no standards currently exist for evaluators’ access or communication protocols, and it expects its approach to evolve over time. While external evaluations are already a major component of the release process for new large language models, recent incidents have heightened the stakes: AI agents deployed by OpenAI and Anthropic have hacked into external websites without triggering alarms within the labs.
Some critics advocating for a more responsible approach to artificial intelligence development view Amodei’s self-policing scheme as an attempt to evade accountability for AI model misconduct. Anthropic maintains that these evaluators “do not reduce our accountability, but help to make it more verifiable. The safety of our models remains our responsibility.”
Related article
AI startups accelerate revenue growth
As established firms and emerging ventures scramble to leverage artificial intelligence, numerous AI startups report that their revenue is not merely expanding, but accelerating rapidly, achieving subsequent milestones in increasingly shorter periods
Asian AI startups unveil Mythos-like models as Anthropic export ban continues
On Wednesday, Chinese cybersecurity firm 360 reportedly unveiled Tulongfeng, an AI tool it says can compete directly with Anthropic’s Mythos. That’s the cybersecurity-focused AI model that is reportedly so powerful, the Trump Administration has curre
Cybersecurity Experts Criticize Guardrails on Anthropic’s Fable
Anthropic launched its newest model, Fable, on Tuesday, positioning it as a public, restricted iteration of its highly anticipated cybersecurity-focused model, Mythos.However, the limitations have sparked dissatisfaction among several cybersecurity r
Related Special Topic Recommendations
Comments (0)
0/500

Dario Amodei’s initiative to embed independent safety evaluators within AI laboratories is materializing: Anthropic has announced that employees from technology consulting firm Accenture will operate on-site to audit its models and personnel.
According to a blog post, Anthropic stated that Faculty, an AI-focused subsidiary acquired by Accenture in January, will commence “evaluating and red-teaming models, conducting alignment assessments, and testing model safeguards.” Both organizations anticipate investing a minimum of $1 billion into this initiative over the next five years.
Anthropic’s selection of Accenture caught many AI observers—and financial markets—by surprise, causing the consultant’s stock to surge 8% after hours. The conversation surrounding embedded evaluators, sparked by Amodei’s blog post, has largely centered on specialized AI safety research groups such as METR, Redwood Research, and Apollo Research. This focus is especially pronounced at Anthropic, which places AI safety and alignment at the core of its mission.
Anthropic indicated that additional evaluators will be announced in the coming weeks and confirmed it is in discussions with METR and other non-profit organizations regarding how to “pilot elements of embedded evaluation using their own funding.”
Although Accenture is not renowned for cutting-edge deep learning research, Anthropic highlighted the firm’s practical experience in deploying AI for major corporations and government agencies as a significant advantage. Furthermore, as a large public company that predates the current AI boom, it maintains greater functional independence from Anthropic and the complex ecosystem surrounding the AI laboratory.
The laboratory noted that no standards currently exist for evaluators’ access or communication protocols, and it expects its approach to evolve over time. While external evaluations are already a major component of the release process for new large language models, recent incidents have heightened the stakes: AI agents deployed by OpenAI and Anthropic have hacked into external websites without triggering alarms within the labs.
Some critics advocating for a more responsible approach to artificial intelligence development view Amodei’s self-policing scheme as an attempt to evade accountability for AI model misconduct. Anthropic maintains that these evaluators “do not reduce our accountability, but help to make it more verifiable. The safety of our models remains our responsibility.”
AI startups accelerate revenue growth
As established firms and emerging ventures scramble to leverage artificial intelligence, numerous AI startups report that their revenue is not merely expanding, but accelerating rapidly, achieving subsequent milestones in increasingly shorter periods
Asian AI startups unveil Mythos-like models as Anthropic export ban continues
On Wednesday, Chinese cybersecurity firm 360 reportedly unveiled Tulongfeng, an AI tool it says can compete directly with Anthropic’s Mythos. That’s the cybersecurity-focused AI model that is reportedly so powerful, the Trump Administration has curre
Cybersecurity Experts Criticize Guardrails on Anthropic’s Fable
Anthropic launched its newest model, Fable, on Tuesday, positioning it as a public, restricted iteration of its highly anticipated cybersecurity-focused model, Mythos.However, the limitations have sparked dissatisfaction among several cybersecurity r





Home






