For the last three years, I have been working on building an alternative to current leading web browsers, firefox,chrome, safari,opera and Microsoft edge,it’s called Ares and you can download the beta version for windows here. Most of leading web browsers have something in common, they are either managed by US based companies or funded by chinese investment (opera). As a long time fan of all of them, I noticed the fact that most of users don’t really care about their data safety on a daily basis. When you look at your data value from a strategic and data governance perspective,it changes everything. In fact, your browser represents over 80% of the time and data you spend on internet. Actually, most of the time you spend on your computer is litteraly inside your browser using various kind of services ,search engines etc.
Why is it important to know who owns your browsing data? Because this data includes at least : your history, your habits, your mindset, your behavior, your researchs, your location, your topics of interest, your password, your banking details, the words typed on your keyboard, your patents, the decision you took ,maybe even the sounds surrounding you and access to your webcam. When you think about it closely, the browser is a free software,which makes it easily viral.It’s so deeply in your habits of consumption that you are not even wondering about it, you just use it. Therefore, chatgpt extensions on chrome have been downloaded millions of times and the recent WebGPU release on chromium shows the path for an era of AI powered browsers. Today ,Opera and Microsoft edge with bing (AI integrated) are already deploying AI capacities for their users. Artificial intelligence is the future and Europe should be part of it, not only as the regulator with tools like GDPR but also as a creator. Europe should take advantage of a powerful local smes ecosystem and open sources technologies bringing more transparency and community driven growth to AI R&D. European leaders can either decide that it’s not important to have a role into the AI race, or decide to join that race and make serious investment to European companies and researchers building AI solutions.
For the past 20 years, there has been a kind of status quo over the fact that there’s no major alternative to US based browsers. I think this is wrong because it gives too much power to a few companies which despite all good intentions have to respect local legislation. In the case of US based browsers,it’s called “Patriot Act” and us legislation on AI strategic algorithms exportations. The current US legislation gives extensive powers to US government over data managed by US entities and less power to privacy rights principles, as simple as it is. If you are a basic user, it may not sound important to you (actually, it should be).If you are a company, an academic insititution, a public service, or anyone willing to keep his data safe and benefit from AI power without sharing data to a company under a permissive legislation, we have a solution for you, we called it Ares browser. It’s an open source browser based on chromium (as most of the browsers today) which guarantees a standard of quality with 4 MAJOR differences:
1 – Our startup, Faraday, is based in France and Belgium.As an organization based in Europe, we can guarantee to our users that their data is protected by GDPR rules and all our research and development efforts are focused on data privacy, data hosting in Europe, openess and cybersecurity. We are supported by the french non profit Wise Society and part of Nvidia inception program . Our ecosystem of partners gives us extensive access to cloud and GPU capacities for finetuning our models. You can try for free our text to image module called Ares Creative available on https://arescreative.fr based on Stabble Diffusion XL.
2- We believe that artificial intelligence should be open source and accessible to anyone in the world trough a browser without discrimination based on origin,country,level of education and geopolitical interests. Today ,most AI solutions are made available first in western countries or requires complex coding skills/IT tools integration unavailable for the average internet user.In the current situation , a lot of people from latin america, asia,africa,middle east are considered low interest markets by AI leading firms .Therefore, the current leading AI solutions makers are not taking enough developing countries reality into consideration. For example, in my homeland, Cameroon, people don’t have access to chatgpt. Some major AI services providers require an sms confirmation, assuming that everybody has a phone… well, that’s not the case. A lof of people, actually, over a billion people on this planet, don’t have a personal mobile phone, but they can have access to a computer at school, university etc. It may sound strange from a western perspective, but a simple sms request can be discriminative for billions of people with low income.
This is the opposite of our inclusive vision,which implies that AI can empower people and reduce inequalities, create new sources of incomes for content creators from Africa for example.
3– We are a company with social impact values at the core of it’s vision. We reinvest part of our profit to community building, education projects in Africa and entrepreneurship. All the data centers suppliers we are working with are certified to be carbon neutral , we do care about environment impact of technology. It’s also against to handle the data bias in training the AI models. Most of people on the planet are not white and living in a western country, that’s a simple fact, but most AI models are trained on data from these countries with these countries perspective and culture. AI shoudl be more inclusive and datasets should respect all racial,cultural and social backgrounds as much as possible. If you write the prompt “beautiful woman” in most text-to-image models, including ourselves, today, you have 5 times more chances to get a white than a black woman. This should change with more inclusive datasets.
4– We believe in OFFLINE Artificial intelligence. Again,it’s important to remind what’s the reality for over 3 billion people. Internet is not yet accessible to everyone ,everytime ,everywhere with almost unlimited data. Unlimited and affordable internet access is true in Europe, North America and rich countries, it’s not true at all for the huge majority of people on the planet. The need of cloud powered AI and the fight to get the most powerful AI of the week, is not the most important from our perspetive. In fact, we believe the current state of open sources AI tools can already do a lot. African entrepreneurs can take the lead in the AI field. To name a few tools we are working on to build Ai capacities within our browser:
- *GPT4ALL ,which proves it’s possible to have a large language model running locally without internet and have been used by millions of people within 3 months.
- * Hugging Chat , a tool allowing any company like us to build it’s own AI chat model with a strong UI base.
- * Stabble Studio, just released as open source by Stabble diffusion for text to image tools creators.
- *Open assistant , a dataset built by the community.
- *Vicuna, giving a quality close to 90% of what you can get with chatgpt.
- * WebLLM : a project taking advantage of webgpu for llm integration into browsers.
We are currently building several additional AI powered modules and cybersecurity features, to know more, follow us on social media on twitter “Ares Creative” and on Facebook “Faraday lab”. Our github repo : https://github.com/Faraday-Lab
If you want to support the project or get in touch with us, feel free – firstname.lastname@example.org . Don’t hesitate to download the beta on https://faradaylab.fr , it’s still at early stage and our most important modules are not availabe yet to the public.