Unleash the democratising power of software developers and the community as a fundamental first step to achieve the development of Secure Artificial Intelligence.

Preamble

Free software has played an important role in the development and evolution of the digital age. The world of computing would not be the same if it were not for the principles and philosophy that govern it.

These principles grant 4 key freedoms: use, study, distribution, and improvement of the software. However, the greatest value emerges from collaborative work, from learning that provides an open source, inspiration, and co-operation of the creative process.

Similarly, open source has also played an important role. Both have been concerned with establishing a regulatory framework for dealing with proprietary software. Such a framework is clearly defined through the different types of software licences that exist. For example, the GPL (GNU General Public Licence) requires inheritance of that licence in new creations or changes based on a GPL-type licence. It is a mature practise to be able to clearly show the terms and conditions under which a particular type of software is released. Similarly, there is an established culture for creative work that does not necessarily refer to software development, such as Creative Commons licenses.

All of that is great because free and open source software licences have facilitated the development and accelerated growth of the digital age.

However, times have changed since Richard Stallman announced the principles of Free Software. Today the rapid proliferation of Artificial Intelligence driven by the democratization of Data Science, is just beginning, and new risks are coming to light.

The problem

With the evolution of Artificial Intelligence (AI), the rapid growth of AI applications and advances in this field, many people, organizations, small and large companies are beginning to worry about the risk associated with the explosion of Artificial Intelligence.

Several are the cases reported to Artificial Intelligence as out of control, so regulating its development is imperative according to the criteria of various analysts.

Recently, in the International Joint Conferences on Artificial Intelligence Organization (IJCAI) some 2,400 people and 160 companies from 90 countries signed an agreement not to develop weapons with AI, recording their commitment and concern on this issue, however, there are unique opportunities to prevail and extend this type of agreement globally.

The opportunity

Considering that the software developer is at the center of the creative and transformative activity of the digital age, that his creative work will define the future of Artificial Intelligence, that the community of developers is characterized by its humanistic sense of collaboration and social commitment, then, I consider it opportune and necessary to give the tools so that “the developer or creator”, together with the community, can play a more active role through a new form of licensing of their creative work.

When, in the early 1980s, Richard Stallman laid the foundation for free software, the risks associated with AI development were not an issue of concern in the developer community. Had it been, free software activists would likely have considered a responsible use clause in their software licence terms. After all, who would want their source code, published under the terms and conditions of an MIT license, such as, to end up being used against humanity? This question may have a clear answer in most of the community, however, the introduction of a new licence poses several challenges.

The proposal

Definition of the SRU licence in the legal and functional framework. SRU: *New Socially Responsible Use Licence”

Challenges

Considering the mare magnum of licencing options existing in this industry:

  • Is the definition of a new software licence justified?

  • How to make it without adding more confusion to the existing problem of licence proliferation, without duplicating reestablished concepts or interfering with existing ones?

  • How to apply it effectively and without social exclusion?

  • How would you give to the development of Secure Artificial Intelligence?

These are some challenges that a new type of licence will have to solve.

Motivation

The main motivation of this work is to take small steps to regulate the development of Artificial Intelligence and to give to the development and socially responsible use of free and open source software, facilitating the establishment of human condition and dignity above all else.

The community has the power to limit the use of their creations and thus keep the constructive and social spirit that has motivated them in their development.

It is a small step but a solid and determining step at the moment it acquires a legal connotation.

What is expected of the community?

The first step towards achieving Secure Artificial Intelligence begins with the will of its creator: “The Man”.

The community is expected to be aware of the power of change it can offer by adopting a new type of licencing for Socially Responsible Use of Software Licences.

How to do it

  • Offering a new licencing alternative that is practical and current,” according to The Times. Establishing a contractual framework based on the existing principles of free and open source software and that includes a clause that obliges the receiving party to adopt a commitment to socially responsible use.

  • By building, consolidating and enriching the relationship of existing free and open source software communities.

  • To help with the integration, practically the implementation of an SRU licence would be to consider it as an annex to the existing licenses, and where it is clearly stipulated the obligation of a Socially Responsible Use.

  • Using existing technologies to ease adoption.

In the next blog post I will continue to address this issue.

Cite this post as:

Hernández Barroso, Raciel. «A Necessary Step for Safe Ai». Blog, 5 de julio de 2019. https://racherb.github.io/2019-07-05-the-first-step-to-a-safe-ai/.

Share and Comment 👇