G
Blog
October 6, 2021
2 min read

How did the idea of Giskard emerge? #7 👮‍♀️ Regulation

Biases in AI / ML algorithms are avoidable. Regulation will push companies to invest in mitigation strategies.

Frances Haugen testifying at the US Senate
Alex Combessie
Frances Haugen testifying at the US Senate
Frances Haugen testifying at the US Senate

It is also about the need for regulation.

There are several reasons why I think regulation would be good news for our profession.

1️. Regulation can reduce the risks of faulty/biased AI algorithms

There have been too many scandals related to AI in the last three years. The last one comes from Facebook (no surprises).

If you listen to Frances Haugen, the whistleblower who dared to reveal the truth, one sentence rings a bell:

“These problems are solvable.” 👀

Indeed, from a technical standpoint, biases/defects in AI algorithms are avoidable. What we are missing is regulation to push companies such as Facebook to invest in mitigation strategies.

2. Regulation can restore trust between AI professionals and the public

When I started working in the AI industry seven years ago, most people around me were excited about the prospects of AI to improve our lives.

Today, that wave of positivism is fading. More and more people I talk to outside of our industry feel defiant about my work. More and more AI professionals I know are thinking about changing careers. 😔

I am still optimistic: AI can be a force of good. Back in 2017, I was writing about the need for the Hippocratic Oath of data scientists.

Regulation can push in the right direction and restore trust with the users, the people whom we serve.

3️. Regulation can improve the quality of our work as AI professionals

Will regulation make the work of AI professionals less productive and less fun? I believe quite the contrary.

You can look at other professions and how they reacted to regulation. Civil engineers continue to innovate and create evermore innovative structures. Investment professionals continue to find novel ways to finance innovation & growth in our economies.

You can also look at software engineering. It is not highly regulated. Yet, software professionals have developed an impressive body of self-regulation. Think QA, Agile, TDD, etc.

Perhaps regulation is an opportunity for us to develop “AI craftsmanship”?

The good news is that regulations are coming! Find out more about the:

🇪🇺 European regulation

🇺🇸 USA regulation

Integrate | Scan | Test | Automate

Giskard: Testing & evaluation framework for LLMs and AI models

Automatic LLM testing
Protect agaisnt AI risks
Evaluate RAG applications
Ensure compliance

How did the idea of Giskard emerge? #7 👮‍♀️ Regulation

Biases in AI / ML algorithms are avoidable. Regulation will push companies to invest in mitigation strategies.
Learn more

It is also about the need for regulation.

There are several reasons why I think regulation would be good news for our profession.

1️. Regulation can reduce the risks of faulty/biased AI algorithms

There have been too many scandals related to AI in the last three years. The last one comes from Facebook (no surprises).

If you listen to Frances Haugen, the whistleblower who dared to reveal the truth, one sentence rings a bell:

“These problems are solvable.” 👀

Indeed, from a technical standpoint, biases/defects in AI algorithms are avoidable. What we are missing is regulation to push companies such as Facebook to invest in mitigation strategies.

2. Regulation can restore trust between AI professionals and the public

When I started working in the AI industry seven years ago, most people around me were excited about the prospects of AI to improve our lives.

Today, that wave of positivism is fading. More and more people I talk to outside of our industry feel defiant about my work. More and more AI professionals I know are thinking about changing careers. 😔

I am still optimistic: AI can be a force of good. Back in 2017, I was writing about the need for the Hippocratic Oath of data scientists.

Regulation can push in the right direction and restore trust with the users, the people whom we serve.

3️. Regulation can improve the quality of our work as AI professionals

Will regulation make the work of AI professionals less productive and less fun? I believe quite the contrary.

You can look at other professions and how they reacted to regulation. Civil engineers continue to innovate and create evermore innovative structures. Investment professionals continue to find novel ways to finance innovation & growth in our economies.

You can also look at software engineering. It is not highly regulated. Yet, software professionals have developed an impressive body of self-regulation. Think QA, Agile, TDD, etc.

Perhaps regulation is an opportunity for us to develop “AI craftsmanship”?

The good news is that regulations are coming! Find out more about the:

🇪🇺 European regulation

🇺🇸 USA regulation

Get Free Content

Download our AI Security Guide and learn everything you need to know about key AI security concepts.