QUESTIONI GENERALI SUL DIRITTO DELL’INFORMATICA – Pagina 2

Ancora su intelligenza artificiale e proprietà intellettuale: indagine dell’ufficio USA

Il tema dei rapporti tra proprietà intellettuale (PI) e intelligenza artificiale (AI) è sempre più al centro dell’attenzione.

L’ufficio brevetti e marchi statunitense (USPTO) ha appena pubblicato i dati di un’indagine (request for comments, RFC) su AI e diritti di PI (ci son state 99 risposte, v. Appendix I, da parte di enti ma anche di individuals) : USPTO’s report “Public Views on AI and IP Policy”, ottobre 2020 (prendo la notizia dal post 12.10.2020 di Eleonora Rosati/Bertrand Sautier in ipkat).

Il report (id est, le risposte riferite) è alquanto interessante. Segnalo:

1° – INVENZIONI

le risposte non ritengono necessarie modifiche al diritto brevettuale: alla domanda 3 (<Do current patent laws and regulations regarding inventorship need to be revised to take into account inventions where an entity or entities other than a natural person contributed to the conception of an invention?>), la maggiornza delle risposte < reflected the view that there is no need for revising patent laws and regulations on inventorship to account for inventions in which an entity or entities other than a natural person contributed to the conception of an invention.>, p. 5. Alla domanda 4 (<Should an entity or entities other than a natural person, or company to which a natural person assigns an invention, be able to own a patent on the AI invention? For example: Should a company who trains the artificial intelligence process that creates the invention be able to be an owner?>) , la larga maggiorahza ha detto che <no changes should be necessary to the current U.S. law—that only a natural person or a company, via assignment, should be considered the owner of a patent or an invention. However, a minority of responses stated that while inventorship and ownership rights should not be extended to machines, consideration should be given to expanding ownership to a natural person: (1) who trains an AI process, or (2) who owns/controls an AI system>, p. 7
sulla domanda 10 (<Are there any new forms of intellectual property protections that are needed for AI inventions, such as data protection? Data is a foundational component of AI. Access to data>), le risposte sono invece divise: <Commenters were nearly equally divided between the view that new intellectual property rights were necessary to address AI inventions and the belief that the current U.S. IP framework was adequate to address AI inventions. Generally, however, commenters who did not see the need for new forms of IP rights suggested that developments in AI technology should be monitored to ensure needs were keeping pace with AI technology developments.
The majority of opinions requesting new IP rights focused on the need to protect the data associated with AI, particularly ML. For example, one opinion stated that “companies that collect large amounts of data have a competitive advantage relative to new entrants to the market. There could be a mechanism to provide access to the repositories of data collected by large technology companies such that proprietary rights to the data are protected but new market entrants and others can use such data to train and develop their AI.”>, p. 15

2 – ALTRI DIRITTI DI PI

domanda 1: la creazione da parte di AI è proteggibile come diritto di autore? No de iure condito e pure de iure condendo: <The vast majority of commenters acknowledged that existing law does not permit a non-human to be an author (outside of the work-for-hire doctrine, which creates a legal fiction for non-human employers to be authors under certain circumstances); they also responded that this should remain the law. One comment stated: “A work produced by an AI algorithm or process, without intervention of a natural person contributing expression to the resulting works, does not, and should not qualify as a work of authorship protectable under U.S. copyright law.”109 Multiple commenters noted that the rationale for this position is to support legal incentives for humans to create new works.110 Other commenters noted that AI is a tool, similar to other tools that have been used in the past to create works: “Artificial intelligence is a tool, just as much as Photoshop, Garage Band, or any other consumer software in wide use today … the current debate over whether a non-human object or process can be ‘creative’ is not new; the government has long resisted calls to extend authorship to corporations or entities that are not natural humans>, p. 20-21
domanda 2: quale livello di coinvolgimento umano serve allora per la proteggibilità [domanda molto rilevante nella pratica!!] ? Non si può che vederlo caso per caso: <More broadly speaking, commenters’ response to this question either referred back to their response to the first question without comment (stating that human involvement is necessary for copyright protection) or referred back and made some further observations or clarifications, often pointing out that each scenario will require fact-specific, case-by-case consideration. Several commenters raised or reiterated their view that natural persons, for the foreseeable future, will be heavily involved in the use of AI, such as when designing models and algorithms, identifying useful training data and standards, determining how technology will be used, guiding or overriding choices made by algorithms, and selecting which outputs are useful or desirable in some way. The commenters thus predicted that the outputs of AI will be heavily reliant on human creativity>, p. 22.
dom. 7 sull’uso di AI nelle ricerche sui marchi: v. la distinzione tra uso dell’USPTO e uso dei titolari di marchio, p. 31 ss.
dom. 9 sulla protezione dei database, p. 36 ss.: la normativa attuale è adeguata e non c’è bisogno di introdurne una ad hoc come in UE : <Commenters who answered this question mostly found that existing laws are adequate to continue to protect AI-related databases and datasets and that there is no need for reconsidering a sui generis database protection law, such as exists in Europe. Furthermore, one commenter cautioned “that AI technology is developing rapidly and that any laws proposed now could be obsolete by the time they are enacted>, p. 37

Report del Parlamento UE sul nesso tra intelligenza artificiale (AI) e proprietà intellettuale (PI)

è uscito il <REPORT on intellectual property rights for the development of artificial intelligence technologies> (2020/2015(INI)) – A9-0176/2020 del 2 ottobre 2020, approvato dal Parlamento UE (Commissione on Legal Affairs-relatore Stéphane Séjourné).

Non ci sono grandi novità : ripercorre le principali preoccupazioni e/o esigenze, che chi si interessa di AI è ormai abituato a leggere.

Riporto alcuni passi dalla MOTION FOR A EUROPEAN PARLIAMENT RESOLUTION, p. 3 ss:

nota che i documenti della Commissione dello scorso anno sul tema dell’AI (v. mio post 20.02.2020) non tenevano conto della PI: <notes, however, that the issue of the protection of IPRs in the context of the development of AI and related technologies has not been addressed by the Commission, despite the key importance of these rights;>, § 1, p. 6.
eventuale legislazione dovrà essere tramite regolamento , non direttiva, § 3.
sullo streaming rileva <the importance of streaming services being transparent and responsible in their use of algorithms, so that access to cultural and creative content in various forms and different languages as well as impartial access to European works can be better guaranteed;>, § 8
raccomanda un approccio settoriale e tipologico per la PI, § \0.
circa l’attuazione/enforcement, <acknowledges the potential of AI technologies to improve the enforcement of IPRs, notwithstanding the need for human verification and review, especially where legal consequences are concerned>, § 11;
sui non-personal data ,<is worried about the possibility of mass manipulation of citizens being used to destabilise democracies and calls for increased awareness-raising and media literacy as well as for urgently needed AI technologies to be made available to verify facts and information>, § 18; e osserva che <AI technologies could be useful in the context of IPR enforcement, but would require human review and a guarantee that any AI-driven decision-making systems are fully transparent; stresses that any future AI regime may not circumvent possible requirements for open source technology in public tenders or prevent the interconnectivity of digital services>, § 18, ed ancora: <notes that AI systems are software-based and rely on statistical models, which may include errors; stresses that AI-generated output must not be discriminatory and that one of the most efficient ways of reducing bias in AI systems is to ensure – to the extent possible under Union law – that the maximum amount of non-personal data is available for training purposes and machine learning; calls on the Commission to reflect on the use of public domain data for such purposes>, § 18.

Dal seguente EXPLANATORY STATEMENT, p. 12-13:

le domande di brevetto relato alla AI presso l’EPO sono più che triplicate in dieci anni;
AI è usata ad es. per la ricerca dello stato dell’arte;
rivalutare la PI alla luce dell’AI costituisce una priorità per le UE.

Intelligenza artificiale e machine learning: un ottimo sunto delle relative questioni da parte del Parlamento UK

Il Parlamento britannico pubblica un post riassuntivo delle principali caratteristiche dei fenomeni dell’intelligenza artificale (AI) e machine learning (ML): v. POSTNOTE, n° 633, October 2020 INTERPRETABLE MACHINE LEARNING.

Con la consueta chiarezza e precisione che contraddistinguono la comunicazione divulgativa nella cultura anglosassone.

Riporto solo i concetti di IA e ML (v. Box 1):

<< Artificial intelligence (AI) – There is no universally agreed definition of AI. It is defined in the Industrial Strategy as “technologies with the ability to perform tasks that would otherwise require human intelligence, such as visual perception, speech recognition, and language translation”. AI is useful for identifying patterns in large sets of data and making predictions.

Machine learning (ML) – ML is a branch of AI that allows a system to learn and improve from examples without all its instructions being explicitly programmed. An MLsystem is trained to carry out atask by analysing large amounts of training data and building a model that it can use toprocess future data, extrapolating its knowledge to unfamiliar situations. Applications of ML include virtual assistants (such as Alexa), product recommendation systems, and facial recognition. There is a range of ML techniques, but many experts attribute recent advances to developments in deep learning:

1) artificial neural networks (ANNs).Type of ML that have a designinspiredbythe way neurons transmit information in the human brain.17Multiple data processing units (nodes) are connected in layers, with the outputs of a previous layer used as inputs for the next.

2) deep learning (DL). Variation of ANNs. Uses a greater number of layers of artificial neurons to solve more difficult problems.16DL advances have improved areas such as voice and image recognition >>.

Il post si sofferma alquanto sulla “interpretabilità”. Tema importante, nei limiti in cui una decisione venga presa sulla base di AI/ML (diverrà dunque sempre più importante): il destinatario, per esaminarne la correttezza e valutarne l’eventuale impugnabilità, deve infatti senza troppa fatica comprenderne la motivazione.

Si legge ad es. <<Some stakeholders have said that ML that is not inherently interpretable should not be used in applications that could have a significant impact on an individual’s life (for example, in criminal justice decisions). The ICO and Alan Turing Institute have recommended that organisations prioritise using systems that use interpretable ML methods if possible, particularly for applications that have a potentially high impact on a person or are safety critical>> (p. 3).

Non è però chiaro perchè l’interpretability debba essere perseguita solo nelle decisioni più importanti e (a contrario) perchè si possa invece lasciare nell’oscuro totale il destinatario in quelle meno importanti (come distinguere, poi, le prime dalle seconde?).