To view the original post, please click here.
Sweden-based Anch.ai, a startup that offers a platform to accelerate responsible AI adoption within enterprises, today announced it has raised seed funding from Benhamou Global Ventures (BGV). The company plans to use the round, which also saw participation from Terrain Invest, Magnus Rausing, Kent Janér, and Fredrik Andersson, to further develop its horizontally integrated solution for ethical AI governance and draw more customers.
For any enterprise dabbling with AI, the biggest challenge today is to ensure speedy development of the project while complying with GDPR and the proposed EU regulations (that could lead to fines of up to 6% of global turnover) by avoiding ethical AI risks such as unintended bias, privacy intrusion, and social exclusion. Any effort in this direction not only saves money and prevents regulatory hurdles, but also helps towards broader market acceptance of the solution and gaining the trust of all stakeholders.
But, the problem is, if an enterprise fails to zero in on an AI issue, it cannot even think of resolving it and ensuring compliance. McKinsey’s 2021 State of AI report notes that companies lack the capacity to address the full range of AI risks they face, with many even being unclear on the extent of their risk exposure or the harm their AI could cause to society and individuals.
Ethical AI Governance Platform
To solve these challenges, Anch.ai has developed an all-in-one governance solution that allows enterprises to screen, assess, mitigate, audit, and report ethical AI risks. The company was founded in 2018 as a research and consulting think tank that worked to determine the ethical and societal risks of AI. However, a year ago, it pivoted to build a more comprehensive offering.
“AI is self-scaling and self-learning, with insufficient governance to control for ethical and societal risk. If unaddressed, unintended ethical and legal breaches, privacy violations, discrimination, and other pitfalls can lead to financial losses, legal sanctions, and damage to an organization’s reputation. A critical part of that is credible third-party validation, and anch.AI can provide that sort of ethical insurance for organizations,” Anna Felländer, Anch.AI’s founder, told Venturebeat.
“The platform performs a cross-functional responsibility to drive an organization’s orchestration of ethical AI implementation. It is meant to provide the best balance between speed, choice, and due diligence in implementing AI,” she added.
Available on a subscription model, Anch.ai’s platform provides business teams with a self-assessment dashboard that shows a detailed profile of their AI project. Organizations can use this AI-driven tool to screen and visualize specific projects for ethical pitfalls, root causes, and vulnerabilities such as data bias or bias of the creator. Once the pitfalls are identified, the solution uses prediction and recommendation tools to go into detail and assess what ethical AI risks you are exposed to (like privacy intrusion or discrimination), the potential degree of impact, and how to mitigate them.
“Our gap analysis feature detects and measures potential deviations from your organizational values and guidelines in your AI solutions,” the company says on its website. After assessment, the platform provides a range of mitigation tools to solve the problem at hand and ensure compliance.
“The user is provided with a list of actions tailored to their AI use case which can be undertaken to reduce their level of risk. In addition, based on the type and degree of risk, users are alerted to use particular mitigation tools. For example, a high risk of discrimination would alert the user and guide them through the fairness tool to help in better understanding and mitigating this risk,” Felländer added.
The platform’s tooling has been employed by more than 20 enterprises, which includes global biopharmaceutical, retail, and telecommunication companies as well as national agencies across the Nordic region.
Competitors
While several companies help with best practices to prevent AI risks from occurring in the first place, many have also entered the niche space for follow-up AI risk assessment and mitigation. This includes players such as Fiddler AI, Credo AI, and Aporia.
“There are several responsible AI offerings ranging from large enterprises to startups, but they mainly work from a siloed tech/AI perspective. This means that they miss out on the legal/compliance and business perspective of exposure to ethical risks. On one coherent platform, anch.AI helps tech, legal, and leadership teams ensure that their organization is compliant to regulation (both existing and upcoming) and align with their values,” the founder noted.
Moving ahead, Anch.ai plans to build out its offering, with a part of the latest funding going towards strengthening the platform’s insight engine and implementing additional mitigation tools.
“The goal is set for an increase [in] users on the platform, leading to the continuous expansion of the engine… As the EU regulation on AI in 2024 is getting closer, there will be a vast demand to ensure AI ethical compliance. We are focused on meeting the exploding demand for EU regulation on AI compliance. By the end of 2022, we plan to be the perceived market leader in ethical AI governance in North America and the EU,” Felländer said.
