LINDA, the Fennoscandic vision of general artificial intelligence

Co-creative Intelligence is the strategic choice of DIGILE as the covering theme of our next phase research activities. It is a new term to describe how the intelligence of humans and machines will in the future be combined to  create something fascinating. But the future is still very unclear and will depend on many aspects. For example, the development of machine (or artificial) intelligence could continue towards many different directions. Now it is important to discuss about possible futures of  machine intelligence. This blog post is a joint effort of people who are involved in DIGILE activities in many ways. They created a novel concept LINDA, something that I can understand only partially and something I cannot totally agree with. Still it could be the right direction to take. I hope to see many novel approaches to pop up in the near future. Then we can start to compare them and slowly find desirable future directions. With this forewords I thank the authors of this weeks blog post (see the list below the post).

Pauli Kuosmanen

***

Currently some of the best Finnish scientists and industry professionals are compiling strategic research and innovation agendas based on co-creative intelligence (CCI). The word “strategic” is crucial here, as funding will be available only for strategic research ventures beyond the short time development horizon. Accordingly, we need a bold vision from which strategies can be derived. Without such vision, potential strategies are likely to address only short term needs or companies’ “here and now” demands. In any event, the vision is the most important input for our writing. The vision we propose is called LINDA, and in short, LINDA is the Fennoscandic – or Nordic – equivalent of IBM’s Watson. Like all visions, LINDA is values and virtues based, in her case purposely Fennoscandic. She is conceptual and abstract enough, so that we still think out of the box, but she is also “personal” enough, in order to provide concrete guidance for machine learning developers, robot researchers or general artificial intelligence scientists. Consequentially, LINDA will enable us to compile strategic agendas because she is a source of inspiration – she will guide us in our research and innovation ideas, through the motto “what would LINDA think about this?”

Conceiving LINDA

We came up with the idea of LINDA at an CCI ideation workshop on 8 April 2015, in a group of broadly diverse researchers and industry professionals. It was clear from the start that LINDA needed to be a long term vision, be more humane than other general Artificial Intelligence (AI) concepts, and have Finnish values embedded. The name LINDA comes from Estonian mythology, where the heroic Linda is the mother of the main hero of an epic. In principle, LINDA can be called “robust and beneficial artificial intelligence”, as in the recent (2015) “Research priorities for robust and beneficial artificial intelligence” paper, signed by hundreds of AI scientists and other public persons, such as Stephen Hawking or Elon Musk.

So what is LINDA like – what are her values?

Her “personality” is strongly different from IBM’s Watson. Whereas Watson is an “individual”, built for competition, LINDA is rather a community, built for caring. She is social and her co-creation mechanisms are embedded, whereas Watson is distant and data oriented. Consequently, LINDA is consensus seeking and provides feedback, whereas Watson confronts with statements. LINDA represents her values, virtues and concerns, Watson presents in a window dressing fashion. And where Watson goes for simple answers (“I learned cooking – now I propose novel meals”), LINDA co-creates by contemplating the “right” questions. In a way, her questions could be good inputs for Watson. And this is one of the key virtues of LINDA – even if she is, by purpose, different than Watson, she is AI that is integrative and strongly networked. LINDA will always try to collaborate first, and only competes if all else fails, while maintaining a positive and joyful mood.

And what about LINDA’s virtues?

It is needless to mention that LINDA has “sisu” – her perseverance and stamina would impress even Tarja Halonen. At the same time her international understanding and her mediator skills exceed those of Nobel-prize winner Martti Ahtisaari. Put into a vehicle, LINDA could drive faster than Kimi Räikkönen, however one of her core values is sustainability, and thus she prefers to drive in dense flocks of same-speed cars. Her cognitive abilities in visuals, haptics and 3D (“printing”) generation make her the preferred co-creation partner for important and/or growing demographic groups, such as entrepreneurs, intelligent services developers, and the elderly. When she talks, her voice is inspired by Lena Olin, her creativity exceeds all Marimekko designs and her level of care delights senior citizens. Still, LINDA is an inspired teacher, drawing from the latest research in gaming, being easily approachable and egalitarian, and at any moment practicing what she “preaches”. Moreover she does not resist saying “en tiedä” if she does not know an answer and thus invites humans to educate her. What is most important however, is that LINDA can be trusted at every moment, that her integrity will always stay intact and that she is one hundred per cent reliable.

Is LINDA one or many?

Based on LINDA’s value and virtues, especially on her reliability and trustworthiness, it is easier to imagine LINDA as a group of intelligent experts than one “lone player”. Her role model is not James Bond, rather the diverse characters of a successful start-up. Being deeply egalitarian, LINDA accepts of course humans, and also other, slightly different (for instance in their cognitive abilities) “LINDAs” and even Watson by her side – for her, all those are rather sources of diverse inspiration than competition.

Will LINDA become real?

It is of crucial importance to understand that LINDA is conceptual, she is a vision. She guides us in writing strategies and inspires us in compiling research and innovation agendas for Finland and beyond. LINDA will mature, this text is just a first “draft” of LINDA. If some researchers and developers are bold enough to start implement a general AI based on LINDA – great! If not, we will still perform research and develop co-creative intelligence based services that will push Finland to the leading edge of AI based offerings. LINDA can promise that.

What next?

We feel that LINDA is inspirational for all CCI SR(I)As. We also feel that LINDA is only one of the “stories about the future” that we need for writing excellent research and innovation agendas. Please contact Tiina Kymäläinen (tiina.kymalainen@vtt.fi) and Raphael Giesecke (Raphael.Giesecke@aalto.fi) in case you have good ideas for visionary stories. Meanwhile, we invite all of you to comment and “mature” our LINDA vision further, after all one of her core values is consensus.

Raphael Giesecke

Pia Erkinheimo

Heini Ikävalko

Antti Sipilä

Further WS participants

Leave a Reply