Akamai launches firewall to protect enterprise AI from cyber threats

Akamai has introduced Firewall for AI, a new solution aimed at providing security for AI applications against emerging cyber threats.

According to findings in Akamai's recent State of the Internet report titled 'State of Apps and API Security 2025: How AI is Shifting the Digital Terrain', the increased uptake of AI tools and AI-powered APIs in enterprises is contributing to a more hazardous cyber threat environment.

The company states that traditional web application firewalls are not equipped to deal with threats posed by contemporary AI systems, including large language models (LLMs) and agentic AI, which are being rapidly deployed across sectors. As these innovations proliferate, organisations are increasingly exposed to risks such as adversarial attacks, model extraction, API abuse, and large-scale data scraping.

Rupesh Chokshi, Senior Vice President and General Manager, Application Security at Akamai, said: "Traditional security solutions do not stop AI threats. As enterprises embrace the AI era to unlock new levels of productivity, AI security must be a key enabler. Firewall for AI is a game changer for any enterprise adopting AI. Securing AI applications isn't just about blocking attacks, it's about enabling innovation without compromising security or performance. With Firewall for AI, enterprise models, data, and intellectual property are protected from evolving threats."

The Firewall for AI provides multilayered protection that aims to block adversarial inputs, unauthorised queries, and attempts at large-scale data scraping, with the intent of preventing both model manipulation and data exfiltration. It is built to secure both inbound AI queries and outbound AI responses, which addresses security gaps that may be introduced as generative AI technologies are adopted in enterprise environments.

Among the key features highlighted for the Firewall for AI are multilayered protection to prevent manipulation and data theft, real-time detection of AI-specific threats using adaptive security rules, and compliance measures that ensure AI-generated outputs align with industry and regulatory requirements. The solution offers flexibility in deployment through Akamai edge, REST API, or reverse proxy, allowing integration with a range of existing security setups.

The product also includes proactive risk mitigation through filtering of AI outputs to prevent toxic content, hallucinations, and leaks of unauthorised data. The intent is to support enterprises in maintaining control over sensitive information handled by their AI systems.

Akamai is supplementing this launch with API LLM Discovery, an enhancement to its API Security platform, which will automatically uncover generative AI and LLM API endpoints, categorise them, and update security policies on an ongoing basis to help prevent unauthorised access and potential abuse.

These newly announced capabilities are designed to provide businesses with tools required to secure enterprise models, data, and intellectual property as AI adoption continues to expand, and as regulatory and compliance requirements evolve in tandem with the technology.